{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3a11ed1a63976341a88cb2ba", "attempt_id": "be50a1530107", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:27.850031Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 484.9111660150811, "ttft_ms": 115.0154999922961, "decode_ms": 664.3899170449004}, "total_wall_ms": 25551.31404194981, "ttft_ms": 115.0154999922961, "input_usage": {"lookups": 0}, "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": "WG-14H-1182", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "e454fec67e299085db563490eff5a43753bbeed6d185376c1a5962c124db19eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "84aef325cfbe090ee66e0c77", "attempt_id": "ffd9431b36ea", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:29.135154Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 452.45512505061924, "ttft_ms": 106.79108393378556, "decode_ms": 37.88079204969108}, "total_wall_ms": 24884.161750087515, "ttft_ms": 106.79108393378556, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b12dd1e0ce7d3ea403720a15", "attempt_id": "c9cf93e27b2d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:29.751941Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 439.0273329336196, "ttft_ms": 55.669042048975825, "decode_ms": 49.54074998386204}, "total_wall_ms": 24831.298791919835, "ttft_ms": 55.669042048975825, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e516a283116dc1fbf03a09cc", "attempt_id": "43ac7d3476a5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:30.315186Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 469.7635000338778, "ttft_ms": 55.345708038657904, "decode_ms": 47.224792069755495}, "total_wall_ms": 24859.273832989857, "ttft_ms": 55.345708038657904, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c622d56ea939957be41d9d14", "attempt_id": "38c5d13b714b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:30.906316Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 454.5477079227567, "ttft_ms": 86.18487499188632, "decode_ms": 546.9334999797866}, "total_wall_ms": 25374.652167083696, "ttft_ms": 86.18487499188632, "input_usage": {"lookups": 0}, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "81e4b303da0b2eea5b1cf8bb", "attempt_id": "c475edd30bca", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:32.012517Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 468.55662495363504, "ttft_ms": 89.82437499798834, "decode_ms": 619.5766670862213}, "total_wall_ms": 25464.937667013146, "ttft_ms": 89.82437499798834, "input_usage": {"lookups": 0}, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8b5b7fd17028cedde25a13c4", "attempt_id": "c2739179b8b2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:33.209304Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 458.4500420605764, "ttft_ms": 12.439707992598414, "decode_ms": 12.542209005914629}, "total_wall_ms": 24770.533750066534, "ttft_ms": 12.439707992598414, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fe8b4076e72915aaa6fe7277", "attempt_id": "5497662326e1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:33.713039Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 226.30379104521126, "ttft_ms": 12.376750004477799, "decode_ms": 12.279707938432693}, "total_wall_ms": 24537.99679200165, "ttft_ms": 12.376750004477799, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a43fca3afd9d73eafce8209a", "attempt_id": "c07f0dc5068e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:33.983487Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 348.74816704541445, "ttft_ms": 56.17716698907316, "decode_ms": 41.04004194959998}, "total_wall_ms": 24733.256459003314, "ttft_ms": 56.17716698907316, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0b0e56e2c390185e151dc9e5", "attempt_id": "b881b70f0020", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:34.450688Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 520.626625046134, "ttft_ms": 35.6030419934541, "decode_ms": 42.592625017277896}, "total_wall_ms": 24885.92895900365, "ttft_ms": 35.6030419934541, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "86f5b8f30552a1517b27b410", "attempt_id": "c201ff8cf2ff", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:35.070865Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 403.5457500722259, "ttft_ms": 24.656457942910492, "decode_ms": 180.56783394422382}, "total_wall_ms": 24895.891749998555, "ttft_ms": 24.656457942910492, "input_usage": {"lookups": 0}, "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": "Torsel-5", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "0a2de4a55396e5f7b4d24e988f7452e33fe7e43fd6e05881b80bcb3034d74d37"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b1eae00f6fc431a204791b85", "attempt_id": "6693e9d01e44", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:35.700448Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 397.1793750533834, "ttft_ms": 59.966249973513186, "decode_ms": 434.0757089667022}, "total_wall_ms": 25178.303833003156, "ttft_ms": 59.966249973513186, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1379172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "7d9e6c79de4e797c6177e40b49a810a6363041af98982dbd4243baf4bec5c971"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "637d485fc6c1f6626de28b11", "attempt_id": "11d508692f90", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:36.612660Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 387.364333961159, "ttft_ms": 39.14333297871053, "decode_ms": 154.47762503754348}, "total_wall_ms": 24868.222833028995, "ttft_ms": 39.14333297871053, "input_usage": {"lookups": 0}, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5565680d4c764f228ae70aa8", "attempt_id": "a85c5dd421e4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:37.216559Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 386.89499988686293, "ttft_ms": 54.47279196232557, "decode_ms": 39.77145906537771}, "total_wall_ms": 24768.423666944727, "ttft_ms": 54.47279196232557, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8772a430d83a35e83402c97a", "attempt_id": "30e48f02a7ac", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:04:37.721688Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 24286.299792001955, "hops": {"1": 24286.299792001955}, "receiver_prefill_ms": 521.4135829592124, "ttft_ms": 32.74333290755749, "decode_ms": 32.02041599433869}, "total_wall_ms": 24873.46883409191, "ttft_ms": 32.74333290755749, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "406ddd5c4ba0481ed362a33d", "attempt_id": "ea46cdc45f86", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:01.018481Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 591, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 460.54529095999897, "ttft_ms": 94.85354193020612, "decode_ms": 401.17983298841864}, "total_wall_ms": 47933.05520911235, "ttft_ms": 94.85354193020612, "input_usage": {"lookups": 0}, "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": "WG-14H-1182", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "e454fec67e299085db563490eff5a43753bbeed6d185376c1a5962c124db19eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "acab707a6a737db701ac4d5e", "attempt_id": "a8bba113dcce", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:01.995680Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 393.088125041686, "ttft_ms": 20.958708017133176, "decode_ms": 36.77195799537003}, "total_wall_ms": 47427.477584104054, "ttft_ms": 20.958708017133176, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6f4c90c6768d2ada88271440", "attempt_id": "5bd46028d911", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:02.466587Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 381.1927910428494, "ttft_ms": 62.18999996781349, "decode_ms": 40.530583006329834}, "total_wall_ms": 47460.437667090446, "ttft_ms": 62.18999996781349, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "215d501459244dd3bd870a3e", "attempt_id": "4aee956c11ad", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:02.969665Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 521.6127909952775, "ttft_ms": 36.13154194317758, "decode_ms": 31.923625036142766}, "total_wall_ms": 47566.25762605108, "ttft_ms": 36.13154194317758, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a2fff79d1b3034e8ff49ca82", "attempt_id": "9145914325f8", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:03.579901Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 602, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 380.15895802527666, "ttft_ms": 22.87820796482265, "decode_ms": 297.8287920122966}, "total_wall_ms": 47677.405542111956, "ttft_ms": 22.87820796482265, "input_usage": {"lookups": 0}, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5d1e7f5875856ee808ca53b4", "attempt_id": "567596324474", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:04.300475Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 368.2215000735596, "ttft_ms": 97.3021249519661, "decode_ms": 465.41012497618794}, "total_wall_ms": 47907.49895910267, "ttft_ms": 97.3021249519661, "input_usage": {"lookups": 0}, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dc8f5ffe8899d5b21661ae3c", "attempt_id": "4cef94c52e6e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:05.252982Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 396.66791702620685, "ttft_ms": 32.72033389657736, "decode_ms": 29.850834049284458}, "total_wall_ms": 47435.725416988134, "ttft_ms": 32.72033389657736, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "091fac38b4923973500c047c", "attempt_id": "cdfd1741aeb5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:05.733538Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 592, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 355.7032919488847, "ttft_ms": 44.36204198282212, "decode_ms": 44.88733410835266}, "total_wall_ms": 47421.51575104799, "ttft_ms": 44.36204198282212, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c35ae85cf09cc3491d7945b9", "attempt_id": "3e50dabcfe2a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:06.199059Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 608, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 536.536332918331, "ttft_ms": 24.986625066958368, "decode_ms": 38.64104195963591}, "total_wall_ms": 47576.615084079094, "ttft_ms": 24.986625066958368, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e4cb56dc4a9ee3f606aaf48e", "attempt_id": "79a06eecc856", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:06.818802Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 405.3282920503989, "ttft_ms": 35.68054095376283, "decode_ms": 29.271750012412667}, "total_wall_ms": 47446.781166945584, "ttft_ms": 35.68054095376283, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "db2360f8a31f189dfd32b9bb", "attempt_id": "4d1b3d56d5c3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:07.308406Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 609, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 397.4196669878438, "ttft_ms": 50.495791947469115, "decode_ms": 380.64412504900247}, "total_wall_ms": 47805.15020794701, "ttft_ms": 50.495791947469115, "input_usage": {"lookups": 0}, "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": "Torsel-5", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "0a2de4a55396e5f7b4d24e988f7452e33fe7e43fd6e05881b80bcb3034d74d37"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7b34d9fa8f5a7c33ed2de447", "attempt_id": "99b5acfd85d0", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:08.158155Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 604, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 388.9277089620009, "ttft_ms": 40.615833015181124, "decode_ms": 336.48045896552503}, "total_wall_ms": 47742.544000037014, "ttft_ms": 40.615833015181124, "input_usage": {"lookups": 0}, "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": "19,431,432", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "06fe08c35e6644a2505156a3386f0e4d98a25b828e2efd638d491d710f856103"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2865677dbc443a3507c794f3", "attempt_id": "f1d764d981ba", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:08.945087Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 599, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 342.39766700193286, "ttft_ms": 57.88183305412531, "decode_ms": 357.0904169464484}, "total_wall_ms": 47733.93245902844, "ttft_ms": 57.88183305412531, "input_usage": {"lookups": 0}, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d23fbbc40716a443a5f9f251", "attempt_id": "d451a24f1d38", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:09.723055Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 604, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 384.2522921040654, "ttft_ms": 40.89683399070054, "decode_ms": 26.685790973715484}, "total_wall_ms": 47428.408209001645, "ttft_ms": 40.89683399070054, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c4857727baa442909b3eec6f", "attempt_id": "dd2e33bd21f6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:10.195359Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 605, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46975.75791704003, "hops": {"1": 24286.299792001955, "2": 22689.458125038072}, "receiver_prefill_ms": 386.30116707645357, "ttft_ms": 24.932708009146154, "decode_ms": 35.143917077220976}, "total_wall_ms": 47422.90650005452, "ttft_ms": 24.932708009146154, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f9bff4dcf1f4f230b398d234", "attempt_id": "1966136491e6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:56.054379Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 591, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 353.30812504980713, "ttft_ms": 48.51737502031028, "decode_ms": 616.1804998992011}, "total_wall_ms": 93385.24812599644, "ttft_ms": 48.51737502031028, "input_usage": {"lookups": 0}, "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": "WG-14H-1182", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "e454fec67e299085db563490eff5a43753bbeed6d185376c1a5962c124db19eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9ae072588518922e890940fd", "attempt_id": "e6e3b8aaa207", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:57.094489Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 397.93270791415125, "ttft_ms": 40.13604193460196, "decode_ms": 38.915583048947155}, "total_wall_ms": 92844.13904300891, "ttft_ms": 40.13604193460196, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f1e3357a400a15ef7ae63bc0", "attempt_id": "c5056a0d98c3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:57.592468Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 344.4904169300571, "ttft_ms": 59.74529101513326, "decode_ms": 59.517916990444064}, "total_wall_ms": 92831.09533402603, "ttft_ms": 59.74529101513326, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62fb01ff4b9ee8c9a26457d6", "attempt_id": "453c72f09118", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:58.075833Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 555.9015839826316, "ttft_ms": 34.07291602343321, "decode_ms": 35.348166013136506}, "total_wall_ms": 92992.4737920519, "ttft_ms": 34.07291602343321, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5d65c5c9f86e2fec1a8be65b", "attempt_id": "aef995f2f14c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:58.722752Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 602, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 384.52395889908075, "ttft_ms": 24.857625016011298, "decode_ms": 301.6412500292063}, "total_wall_ms": 93078.35558394436, "ttft_ms": 24.857625016011298, "input_usage": {"lookups": 0}, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6513c18b126046258eb3fc2a", "attempt_id": "56ba69958e2c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:05:59.453443Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 422.4503339501098, "ttft_ms": 58.04333300329745, "decode_ms": 537.247542059049}, "total_wall_ms": 93385.06050093565, "ttft_ms": 58.04333300329745, "input_usage": {"lookups": 0}, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c23e40d3ae2b125a823a4f1d", "attempt_id": "e2ed1e12ab93", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:00.492749Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 402.44679199531674, "ttft_ms": 34.497583052143455, "decode_ms": 31.26662503927946}, "total_wall_ms": 92835.28633404057, "ttft_ms": 34.497583052143455, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4e5abec8e24e7c4109039a22", "attempt_id": "eb4a937920dd", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:00.980757Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 592, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 400.2829999662936, "ttft_ms": 58.98241698741913, "decode_ms": 114.35558390803635}, "total_wall_ms": 92941.04575098027, "ttft_ms": 58.98241698741913, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b4f71c100b4623d1b904cc36", "attempt_id": "0a79548edfa0", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:01.577105Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 608, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 468.03841600194573, "ttft_ms": 41.45112494006753, "decode_ms": 24.36329098418355}, "total_wall_ms": 92901.15787589457, "ttft_ms": 41.45112494006753, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e173957db59fdedeab74b81d", "attempt_id": "884611a7a3ea", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:02.134365Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 387.22312496975064, "ttft_ms": 38.671334041282535, "decode_ms": 25.326875038444996}, "total_wall_ms": 92818.71354300529, "ttft_ms": 38.671334041282535, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "94ed3d3b0eefd390d86bf39c", "attempt_id": "610818c09a3c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:02.606757Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 609, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 341.4249159395695, "ttft_ms": 54.5890829525888, "decode_ms": 399.4072920177132}, "total_wall_ms": 93162.65525098424, "ttft_ms": 54.5890829525888, "input_usage": {"lookups": 0}, "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": "Torsel-5", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "0a2de4a55396e5f7b4d24e988f7452e33fe7e43fd6e05881b80bcb3034d74d37"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e03f68e67c4a195a194f971c", "attempt_id": "9f332e2572f0", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:03.423953Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 604, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 399.37645802274346, "ttft_ms": 30.611374997533858, "decode_ms": 339.46141693741083}, "total_wall_ms": 93136.64912595414, "ttft_ms": 30.611374997533858, "input_usage": {"lookups": 0}, "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": "19,431,432", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "06fe08c35e6644a2505156a3386f0e4d98a25b828e2efd638d491d710f856103"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8b500f96abbd98374cb9d510", "attempt_id": "8115fe0c5274", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:04.213413Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 599, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 410.23570904508233, "ttft_ms": 55.81724992953241, "decode_ms": 335.3547089500353}, "total_wall_ms": 93168.69024990592, "ttft_ms": 55.81724992953241, "input_usage": {"lookups": 0}, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "43bb41e19d13d5da689b4511", "attempt_id": "5712eb2a6bce", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:05.037093Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 604, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 398.16812495701015, "ttft_ms": 23.98420893587172, "decode_ms": 40.15366698149592}, "total_wall_ms": 92829.55120899715, "ttft_ms": 23.98420893587172, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "813da9f1f6eaa5b7e55c655d", "attempt_id": "aefa28795edf", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:05.520398Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 605, "message_tokens": 509, "message_utf8_bytes": 1189, "cache_tensor_bytes": null, "serialized_payload_bytes": 1189, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92366.46920896601, "hops": {"1": 24286.299792001955, "2": 22689.458125038072, "3": 22838.89408397954, "4": 22551.817207946442}, "receiver_prefill_ms": 381.688249995932, "ttft_ms": 20.792166935279965, "decode_ms": 42.287042015232146}, "total_wall_ms": 92811.91579205915, "ttft_ms": 20.792166935279965, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "3": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}, "4": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1189}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "49b54c3c28297dd6f3503e10cc594cea5454825636b39e95a6e8663b6c37ed71", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "04e87861d25b179905a4e06f", "attempt_id": "6890cdb76440", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:29.808163Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 422, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 362.3207500204444, "ttft_ms": 98.55062502902001, "decode_ms": 138.78904096782207}, "total_wall_ms": 24422.308334033005, "ttft_ms": 98.55062502902001, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5419baba672619e4bcb38714", "attempt_id": "a328a5959d81", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:30.429195Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 428, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 297.3768749507144, "ttft_ms": 29.200874967500567, "decode_ms": 40.56829202454537}, "total_wall_ms": 24189.682625001296, "ttft_ms": 29.200874967500567, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e05492a9b2fb73cd965c65e1", "attempt_id": "676e89ec2125", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:30.817551Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 428, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 300.59433402493596, "ttft_ms": 26.32541593629867, "decode_ms": 34.4407910015434}, "total_wall_ms": 24183.6928750854, "ttft_ms": 26.32541593629867, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1ef6e6d18893e8ae4002fab0", "attempt_id": "49b11eec8b6f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:31.200773Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 425, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 256.8584999535233, "ttft_ms": 55.39320805110037, "decode_ms": 56.32087495177984}, "total_wall_ms": 24191.148417070508, "ttft_ms": 55.39320805110037, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cace4ca278f6f84ca2bdacf0", "attempt_id": "f09d1df0cba3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:31.589585Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 433, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 515.9304159460589, "ttft_ms": 22.499334067106247, "decode_ms": 258.33300000522286}, "total_wall_ms": 24619.153375038877, "ttft_ms": 22.499334067106247, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1f67651acfd389c44e0dea8f", "attempt_id": "0cb8f17ea8d8", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:32.405997Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 286.7188330274075, "ttft_ms": 34.683541976846755, "decode_ms": 518.2574590435252}, "total_wall_ms": 24662.0955420658, "ttft_ms": 34.683541976846755, "input_usage": {"lookups": 0}, "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": "BH-71Y-2851", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "38b55d8642902ac72f5987064e67fbe9b766a04dc338df11f6fa45d4cb64179e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d0cb251ad1bd77fd569b1a2a", "attempt_id": "6221db80832d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:33.266316Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 425, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 365.08545803371817, "ttft_ms": 36.74712497740984, "decode_ms": 26.03720803745091}, "total_wall_ms": 24250.171875115484, "ttft_ms": 36.74712497740984, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bdaa76a338a55104becc01cc", "attempt_id": "8db79afc697e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:33.715310Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 423, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 293.0555830243975, "ttft_ms": 37.21866698469967, "decode_ms": 23.052165983244777}, "total_wall_ms": 24175.79245811794, "ttft_ms": 37.21866698469967, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "60de06e244d1bad5c877ddde", "attempt_id": "2e3b1862625d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:34.087861Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 439, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 265.4155410127714, "ttft_ms": 12.05600006505847, "decode_ms": 20.229916903190315}, "total_wall_ms": 24120.500958058983, "ttft_ms": 12.05600006505847, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "033fbc81580c992c67114c4e", "attempt_id": "56cdab00f2e7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:34.407235Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 297.8416249388829, "ttft_ms": 52.83541593234986, "decode_ms": 44.32808409910649}, "total_wall_ms": 24217.459584004246, "ttft_ms": 52.83541593234986, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "025ffb29018d6c3d5ca4f70f", "attempt_id": "cc785c78bae5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:34.824218Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 421.43437499180436, "ttft_ms": 29.29579105693847, "decode_ms": 43.11945789959282}, "total_wall_ms": 24316.150916041806, "ttft_ms": 29.29579105693847, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "12cda46b1cb28e7913ebe830", "attempt_id": "ed8018b20d3a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:35.338497Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 435, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 301.94545804988593, "ttft_ms": 40.38058302830905, "decode_ms": 225.62675003428012}, "total_wall_ms": 24390.33441711217, "ttft_ms": 40.38058302830905, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "25bd71914eba20d20c51662a", "attempt_id": "21f08e9db7e0", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:35.926052Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 430, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 294.6047910954803, "ttft_ms": 55.80033303704113, "decode_ms": 43.82391704712063}, "total_wall_ms": 24216.608417104, "ttft_ms": 55.80033303704113, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "18218fa2b4157b2e12b7fa7f", "attempt_id": "0ffe4d815d57", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:36.343318Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 435, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 421.44245805684477, "ttft_ms": 39.56712502986193, "decode_ms": 34.629249945282936}, "total_wall_ms": 24318.20616708137, "ttft_ms": 39.56712502986193, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "af01484eb193de9b93a198b5", "attempt_id": "a84a2488f731", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:36.859367Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 436, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23821.762875071727, "hops": {"1": 23821.762875071727}, "receiver_prefill_ms": 273.0319999391213, "ttft_ms": 34.433749970048666, "decode_ms": 33.00812502857298}, "total_wall_ms": 24162.84720809199, "ttft_ms": 34.433749970048666, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c3d67a89fbe821887daacb9e", "attempt_id": "c8eb621ba10c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:51.527958Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 386, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 297.9855419835076, "ttft_ms": 23.34825007710606, "decode_ms": 18.62379105295986}, "total_wall_ms": 38469.881250173785, "ttft_ms": 23.34825007710606, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3fb96403096481b35667ef7a", "attempt_id": "8a04bc69af3e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:51.889381Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 392, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 294.2246659658849, "ttft_ms": 59.11029200069606, "decode_ms": 59.215291985310614}, "total_wall_ms": 38542.581083020195, "ttft_ms": 59.11029200069606, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d595dd3bb613f377cac8b78f", "attempt_id": "b1b34d7400d1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:52.323949Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 392, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 418.4186659986153, "ttft_ms": 29.895917046815157, "decode_ms": 40.93625000678003}, "total_wall_ms": 38619.02737512719, "ttft_ms": 29.895917046815157, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1fc8975dfb760e7bd445878a", "attempt_id": "204b4426ef6e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:52.832238Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 389, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 297.63920803088695, "ttft_ms": 37.79916593339294, "decode_ms": 31.47345792967826}, "total_wall_ms": 38496.705833124, "ttft_ms": 37.79916593339294, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd063637cfd0d722214d5b5b", "attempt_id": "52e0f1a6fb57", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:53.218822Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 397, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 262.29304098524153, "ttft_ms": 18.856124952435493, "decode_ms": 423.7513749394566}, "total_wall_ms": 38834.85666709021, "ttft_ms": 18.856124952435493, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "26a16ecb19a1a3c646ef06d3", "attempt_id": "a0f9f3425d0a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:53.942819Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 395, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 363.5200000135228, "ttft_ms": 36.026916932314634, "decode_ms": 24.40633298829198}, "total_wall_ms": 38553.928042179905, "ttft_ms": 36.026916932314634, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "978beecdd2e7113d836e9304", "attempt_id": "8b71364925d6", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:54.389904Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 389, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 289.3412499688566, "ttft_ms": 37.92587493080646, "decode_ms": 24.67966708354652}, "total_wall_ms": 38481.78395815194, "ttft_ms": 37.92587493080646, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0d55dc2eba1837f008fc4581", "attempt_id": "dbd17920234d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:54.763260Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 387, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 265.4585409909487, "ttft_ms": 42.12829191237688, "decode_ms": 58.11837501823902}, "total_wall_ms": 38495.53666613065, "ttft_ms": 42.12829191237688, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dab8995e5717a1a3ce5543b6", "attempt_id": "4cc3e5835140", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:55.148506Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 403, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 301.3335840078071, "ttft_ms": 95.95637500751764, "decode_ms": 142.3891669837758}, "total_wall_ms": 38669.47029216681, "ttft_ms": 95.95637500751764, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fcf55be1b8887781667a9c1a", "attempt_id": "5e3f610e1ec5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:55.707880Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 395, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 274.8489159857854, "ttft_ms": 40.59091699309647, "decode_ms": 30.05095908883959}, "total_wall_ms": 38475.34375009127, "ttft_ms": 40.59091699309647, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c55a4b9511a82311cea799fa", "attempt_id": "826efcf44fe2", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:56.072141Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 404, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 275.8313330123201, "ttft_ms": 34.614874981343746, "decode_ms": 42.0795000391081}, "total_wall_ms": 38482.38037514966, "ttft_ms": 34.614874981343746, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1b773b02582bf2e769fcf494", "attempt_id": "ae032689f6bb", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:56.442988Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 399, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 250.94812491443008, "ttft_ms": 53.80141700152308, "decode_ms": 492.20670794602484}, "total_wall_ms": 38926.9722502213, "ttft_ms": 53.80141700152308, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f33182db81c1173cbdeaac77", "attempt_id": "ee59ff73b729", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:57.260067Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 394, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 271.99029200710356, "ttft_ms": 39.312749984674156, "decode_ms": 27.456958079710603}, "total_wall_ms": 38468.61670806538, "ttft_ms": 39.312749984674156, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4b0581c64fd622bf8dbe9671", "attempt_id": "bdf27247932f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:57.618751Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 399, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 273.0154579039663, "ttft_ms": 23.766207974404097, "decode_ms": 36.11387498676777}, "total_wall_ms": 38462.67445804551, "ttft_ms": 23.766207974404097, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fba840b56894fda6636fadea", "attempt_id": "3223759550a7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:06:57.970424Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 400, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 38129.31487511378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051}, "receiver_prefill_ms": 256.6462910035625, "ttft_ms": 66.87658303417265, "decode_ms": 56.20279198046774}, "total_wall_ms": 38509.63833311107, "ttft_ms": 66.87658303417265, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3cf2d66432ff1e12168f8fca", "attempt_id": "d1ed6f59752b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:26.846186Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 386, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 318.09591699857265, "ttft_ms": 44.30162499193102, "decode_ms": 46.81524995248765}, "total_wall_ms": 67013.65237520076, "ttft_ms": 44.30162499193102, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "81b500dc3f1309b66dccb8c0", "attempt_id": "21743f20c950", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:27.277169Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 392, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 418.16979099530727, "ttft_ms": 29.45399994496256, "decode_ms": 40.45470803976059}, "total_wall_ms": 67092.59875013959, "ttft_ms": 29.45399994496256, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3eb830e78edfdb699825c1c5", "attempt_id": "cea7e96ac79f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:27.786281Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 392, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 301.95774999447167, "ttft_ms": 40.36600003018975, "decode_ms": 23.589416057802737}, "total_wall_ms": 66970.25062516332, "ttft_ms": 40.36600003018975, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0374ab8dded0843ca9bf9dd4", "attempt_id": "19ad05075bd1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:28.172940Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 389, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 273.07487500365824, "ttft_ms": 59.0814589522779, "decode_ms": 57.97587498091161}, "total_wall_ms": 66994.56466711126, "ttft_ms": 59.0814589522779, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bad8cfaae6b03e8d6f22ea86", "attempt_id": "fb8cd4c3c90b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:28.582367Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 397, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 281.8677920149639, "ttft_ms": 95.27991700451821, "decode_ms": 360.9344579745084}, "total_wall_ms": 67342.44683315046, "ttft_ms": 95.27991700451821, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "afa8fb4ad4f2b48adc266df8", "attempt_id": "750b96f60893", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:29.340594Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 395, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 282.4939170386642, "ttft_ms": 33.157541998662055, "decode_ms": 41.24349995981902}, "total_wall_ms": 66961.18625008967, "ttft_ms": 33.157541998662055, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "743937e6adfee9a209ed5a47", "attempt_id": "fec3a26b6c76", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:29.716463Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 389, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 254.59937495179474, "ttft_ms": 19.75479198154062, "decode_ms": 45.507167000323534}, "total_wall_ms": 66924.19600021094, "ttft_ms": 19.75479198154062, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "758129c26d295bc7203353f1", "attempt_id": "85f6cd41cd63", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:30.054529Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 387, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 275.49958403687924, "ttft_ms": 61.372875003144145, "decode_ms": 113.07666706852615}, "total_wall_ms": 67054.34595909901, "ttft_ms": 61.372875003144145, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d6ab33ef506e6da4785075e3", "attempt_id": "7d33a74af6fa", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:30.523829Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 403, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 368.36029100231826, "ttft_ms": 36.712541943416, "decode_ms": 32.97900001052767}, "total_wall_ms": 67042.61741705704, "ttft_ms": 36.712541943416, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "435ec35de8fdd6de704455e5", "attempt_id": "4cf1a5997980", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:30.983479Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 395, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 298.6492500640452, "ttft_ms": 36.734916968271136, "decode_ms": 44.631625059992075}, "total_wall_ms": 66984.32758322451, "ttft_ms": 36.734916968271136, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4572f3ac765a0211b8840da1", "attempt_id": "a2c12eec40f0", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:31.382398Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 404, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 268.28149997163564, "ttft_ms": 52.605042001232505, "decode_ms": 35.88162502273917}, "total_wall_ms": 66961.57054218929, "ttft_ms": 52.605042001232505, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6b9499bcd4f72506b77e60bf", "attempt_id": "a1ee4ac81073", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:31.758599Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 399, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 275.8622500114143, "ttft_ms": 91.51529194787145, "decode_ms": 356.07212502509356}, "total_wall_ms": 67327.97008310445, "ttft_ms": 91.51529194787145, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ab587955560882cbd5ce7484", "attempt_id": "3b2ae9edd0b8", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:32.503470Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 394, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 289.80404208414257, "ttft_ms": 37.13112499099225, "decode_ms": 25.725124985910952}, "total_wall_ms": 66956.94216608535, "ttft_ms": 37.13112499099225, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5296369ae55c56612da02259", "attempt_id": "cbfea3161754", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:32.875701Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 399, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 264.95562493801117, "ttft_ms": 50.35633291117847, "decode_ms": 56.641708011738956}, "total_wall_ms": 66976.58729110844, "ttft_ms": 50.35633291117847, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "774dd471a6948e75e833ee39", "attempt_id": "dc00536d0589", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:33.267196Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 400, "message_tokens": 304, "message_utf8_bytes": 725, "cache_tensor_bytes": null, "serialized_payload_bytes": 725, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 66603.82750013378, "hops": {"1": 23821.762875071727, "2": 14307.552000042051, "3": 14404.276708024554, "4": 14070.235916995443}, "receiver_prefill_ms": 324.3159579578787, "ttft_ms": 133.0089159309864, "decode_ms": 85.63491608947515}, "total_wall_ms": 67147.47316716239, "ttft_ms": 133.0089159309864, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "3": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}, "4": {"facts_total": 64, "facts_retained": 14, "retention_rate": 0.21875, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.125, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 725}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "67f71c88a4a191ee9bab456f246ff715b6530970fc95af245c2614536b1f9c0d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4931891ef17c411143e282f7", "attempt_id": "a8c087242a96", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:57.816427Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 391.00658299867064, "ttft_ms": 34.61070801131427, "decode_ms": 322.20470905303955}, "total_wall_ms": 24727.739749941975, "ttft_ms": 34.61070801131427, "input_usage": {"lookups": 0}, "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": "WG-14H-1182", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "e454fec67e299085db563490eff5a43753bbeed6d185376c1a5962c124db19eb"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "68a73507cc6aa162fe3084d2", "attempt_id": "26ac31509080", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:58.586817Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 467.82716701272875, "ttft_ms": 145.8245839457959, "decode_ms": 39.968500030227005}, "total_wall_ms": 24633.3457500441, "ttft_ms": 145.8245839457959, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e780c6163e9e789703ed988a", "attempt_id": "ea849ad34bcd", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:59.263083Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 389.8612920893356, "ttft_ms": 27.10166701581329, "decode_ms": 44.6985000744462}, "total_wall_ms": 24441.427999990992, "ttft_ms": 27.10166701581329, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1d8e19e4cb811835853baad3", "attempt_id": "06a34f6dd7c3", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:07:59.748146Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 380.6191250914708, "ttft_ms": 20.964083028957248, "decode_ms": 20.86420904379338}, "total_wall_ms": 24402.167416992597, "ttft_ms": 20.964083028957248, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1610f1ac615db9d7bb219b70", "attempt_id": "2d535330e323", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:00.194217Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 409.48787494562566, "ttft_ms": 52.08716704510152, "decode_ms": 501.5622089849785}, "total_wall_ms": 24942.737625096925, "ttft_ms": 52.08716704510152, "input_usage": {"lookups": 0}, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1f32b366564448c17b11f515", "attempt_id": "1af8ccc62d29", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:01.182530Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 397.4362090229988, "ttft_ms": 33.432708005420864, "decode_ms": 376.7092499183491}, "total_wall_ms": 24787.256499985233, "ttft_ms": 33.432708005420864, "input_usage": {"lookups": 0}, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f3e6028ee513670114bf8264", "attempt_id": "bfd5fbb7f2da", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:02.013061Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 540.9586670575663, "ttft_ms": 37.65283292159438, "decode_ms": 29.376916936598718}, "total_wall_ms": 24587.597167119384, "ttft_ms": 37.65283292159438, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bd507904aaebe43a02aefd4b", "attempt_id": "371a62be1fc0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:02.644152Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 387.68154196441174, "ttft_ms": 39.986666990444064, "decode_ms": 26.298459037207067}, "total_wall_ms": 24433.51695802994, "ttft_ms": 39.986666990444064, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "db944d315718d86958606bea", "attempt_id": "8df98e8f55e6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:03.119392Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 358.64304192364216, "ttft_ms": 35.44391691684723, "decode_ms": 63.01991699729115}, "total_wall_ms": 24436.85270799324, "ttft_ms": 35.44391691684723, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c6311247a406f02340442ffa", "attempt_id": "1e4fa5744b6a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:03.597772Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 559.3518749810755, "ttft_ms": 31.410999945364892, "decode_ms": 42.6307920133695}, "total_wall_ms": 24613.05666610133, "ttft_ms": 31.410999945364892, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6013365c1510b59ff8c3f38a", "attempt_id": "a3a37e05b35d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:04.254734Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 387.72283296566457, "ttft_ms": 29.165541054680943, "decode_ms": 192.58320797234774}, "total_wall_ms": 24589.05974996742, "ttft_ms": 29.165541054680943, "input_usage": {"lookups": 0}, "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": "Torsel-5", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "0a2de4a55396e5f7b4d24e988f7452e33fe7e43fd6e05881b80bcb3034d74d37"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6dc46da6c9a2438c92be312c", "attempt_id": "bba1c1992890", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:04.884968Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 341.6317499941215, "ttft_ms": 42.945874971337616, "decode_ms": 475.6929170107469}, "total_wall_ms": 24840.00404109247, "ttft_ms": 42.945874971337616, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1379172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "7d9e6c79de4e797c6177e40b49a810a6363041af98982dbd4243baf4bec5c971"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e76d7988ea1a8aa9af55f48", "attempt_id": "e6894c5f6fb8", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:05.765784Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 392.82700000330806, "ttft_ms": 33.00854202825576, "decode_ms": 138.3604999864474}, "total_wall_ms": 24544.085458037443, "ttft_ms": 33.00854202825576, "input_usage": {"lookups": 0}, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f221bc12e259bde57e7bbb64", "attempt_id": "17a436515a0b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:06.349063Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 364.2572920070961, "ttft_ms": 57.67416697926819, "decode_ms": 42.04425006173551}, "total_wall_ms": 24443.564250017516, "ttft_ms": 57.67416697926819, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "353f6d5326adb919d4ad97ac", "attempt_id": "492940c352df", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:06.833128Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1196, "cache_tensor_bytes": null, "serialized_payload_bytes": 1196, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23978.786875028163, "hops": {"1": 23978.786875028163}, "receiver_prefill_ms": 470.69612494669855, "ttft_ms": 97.23412501625717, "decode_ms": 29.755542054772377}, "total_wall_ms": 24577.265875064768, "ttft_ms": 97.23412501625717, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f811e3d2b4b96fb92f1cc9651f50294bf8a18ad7c33cafac648e5969f236fb1e", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7ba04cfd656e89790f8f39a6", "attempt_id": "38133032a8ec", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:29.076469Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 565, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 403.31441606394947, "ttft_ms": 93.91537494957447, "decode_ms": 133.38787492830306}, "total_wall_ms": 46232.21958405338, "ttft_ms": 93.91537494957447, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8c35bf34df1625099f80d7fb", "attempt_id": "90258bf59e0b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:29.727949Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 571, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 367.03466705512255, "ttft_ms": 37.405582959763706, "decode_ms": 31.54983406420797}, "total_wall_ms": 46037.49337501358, "ttft_ms": 37.405582959763706, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e8fbd75fac3ddc7dcea1f6af", "attempt_id": "746d06ecd7ae", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:30.183799Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 571, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 365.8348750323057, "ttft_ms": 25.32420796342194, "decode_ms": 37.92033297941089}, "total_wall_ms": 46030.54633399006, "ttft_ms": 25.32420796342194, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ab85246b03b00c7139821d1c", "attempt_id": "55127dcff71f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:30.632972Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 568, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 357.23279195372015, "ttft_ms": 55.18312496133149, "decode_ms": 50.56145798880607}, "total_wall_ms": 46064.71333396621, "ttft_ms": 55.18312496133149, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d34f3730c8960b49a97dacf3", "attempt_id": "0e1c5bf1a449", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:31.116925Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 576, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 511.3057499984279, "ttft_ms": 35.30379198491573, "decode_ms": 332.57245796266943}, "total_wall_ms": 46480.70625006221, "ttft_ms": 35.30379198491573, "input_usage": {"lookups": 4}, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a746c610ec4bf63abe0c47ba", "attempt_id": "8b8595104d04", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:32.017147Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 574, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 369.3179580150172, "ttft_ms": 58.444582973606884, "decode_ms": 608.6580420378596}, "total_wall_ms": 46637.9206670681, "ttft_ms": 58.444582973606884, "input_usage": {"lookups": 4}, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "53c0674c2088c76582700146", "attempt_id": "8cb40d60d5bd", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:33.074014Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 568, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 376.7236249987036, "ttft_ms": 30.88629199191928, "decode_ms": 40.62300000805408}, "total_wall_ms": 46049.81520900037, "ttft_ms": 30.88629199191928, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f0b60eb9c3615749aca99d67", "attempt_id": "afddd2468909", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:33.541588Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 566, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 348.9598749438301, "ttft_ms": 55.41720800101757, "decode_ms": 35.81837494857609}, "total_wall_ms": 46042.121208040044, "ttft_ms": 55.41720800101757, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2e721023de317e4074cdbe69", "attempt_id": "b317b6366138", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:34.002851Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 582, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 533.8108340511099, "ttft_ms": 36.33241693023592, "decode_ms": 32.04508405178785}, "total_wall_ms": 46203.751792083494, "ttft_ms": 36.33241693023592, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d90cec963461808a6b5cd488", "attempt_id": "1d2558723b52", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:34.625252Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 574, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 369.0517919603735, "ttft_ms": 39.14820903446525, "decode_ms": 24.52304190956056}, "total_wall_ms": 46034.19512603432, "ttft_ms": 39.14820903446525, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ce11b3ac970dca471e77f0db", "attempt_id": "4a0f5e16fdf2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:35.077253Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 583, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 369.21491706743836, "ttft_ms": 38.46345900092274, "decode_ms": 63.170625013299286}, "total_wall_ms": 46072.42483308073, "ttft_ms": 38.46345900092274, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "63d6af705f0dfa5429753c45", "attempt_id": "26f2aac537ad", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:35.569966Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 578, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 449.49854200240225, "ttft_ms": 142.892999923788, "decode_ms": 338.11433403752744}, "total_wall_ms": 46531.98737511411, "ttft_ms": 142.892999923788, "input_usage": {"lookups": 4}, "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": "14,039,232", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "e36f51d5f0cc4186b84faf74cd8ec4a754f1d19fb316576662b107b680533008"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7d316d32e10e5db91f67c2a1", "attempt_id": "4b72790375d9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:36.521547Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 573, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 372.00204201508313, "ttft_ms": 32.76120906230062, "decode_ms": 110.6048331130296}, "total_wall_ms": 46116.837250068784, "ttft_ms": 32.76120906230062, "input_usage": {"lookups": 4}, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0dca7f2483cdf086a3314439", "attempt_id": "614f6118812c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:37.058113Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 578, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 416.2859170464799, "ttft_ms": 92.98574994318187, "decode_ms": 142.13183394167572}, "total_wall_ms": 46252.961500082165, "ttft_ms": 92.98574994318187, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f76e0f2eb973ec259a0d8b56", "attempt_id": "c38950e0db6e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:08:37.730104Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 579, "message_tokens": 484, "message_utf8_bytes": 1076, "cache_tensor_bytes": null, "serialized_payload_bytes": 1076, "side_information_bytes": 0, "fetched_bytes": 789, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45600.79866705928, "hops": {"1": 23978.786875028163, "2": 21622.011792031117}, "receiver_prefill_ms": 379.8245000652969, "ttft_ms": 32.2311669588089, "decode_ms": 33.74975000042468}, "total_wall_ms": 46047.3206670722, "ttft_ms": 32.2311669588089, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "61f2cf6ef72ffa12b1b50c4e80cd6f6976b44b30bc453a544a8438adf456b813", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "175ece8d72ceff8d9ea2c7ce", "attempt_id": "780cace8fe95", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:19.509147Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 521, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 465.0091660441831, "ttft_ms": 26.796999969519675, "decode_ms": 44.384874985553324}, "total_wall_ms": 87449.41170804668, "ttft_ms": 26.796999969519675, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3d9ccffdc5309f06edd7f378", "attempt_id": "bed8af033797", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:20.072119Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 527, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 354.66712492052466, "ttft_ms": 25.046250084415078, "decode_ms": 38.4860000340268}, "total_wall_ms": 87331.37649996206, "ttft_ms": 25.046250084415078, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9b27483d77b2e3531245dded", "attempt_id": "2e78c705d3c2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:20.511485Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 527, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 338.99199997540563, "ttft_ms": 57.25216690916568, "decode_ms": 57.745959027670324}, "total_wall_ms": 87367.1367081115, "ttft_ms": 57.25216690916568, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aa3ba5458c817398a3359120", "attempt_id": "bef9aa50b8a9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:20.986143Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 524, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 512.6552499132231, "ttft_ms": 36.28137498162687, "decode_ms": 37.45379205793142}, "total_wall_ms": 87499.7093329439, "ttft_ms": 36.28137498162687, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f8712176897048bbe3638ef0", "attempt_id": "083147f88b33", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:21.591671Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 532, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 358.70591702405363, "ttft_ms": 25.25158307980746, "decode_ms": 291.58545890823007}, "total_wall_ms": 87588.62566703465, "ttft_ms": 25.25158307980746, "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": "6,889,172", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "76875b24cee85c9daa41a3b2716f4617e3b85d1aa6a8a5fd34f349dcf4ae3d7e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "180380594eccfcc469bfee87", "attempt_id": "0bc96ef06e0d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:22.286716Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 530, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 348.67691702675074, "ttft_ms": 41.11158300656825, "decode_ms": 542.3874999396503}, "total_wall_ms": 87845.20970808808, "ttft_ms": 41.11158300656825, "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": "SP-38G-4217", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "e643c21a78559b96fd65325d9ce779b7b8fd6a47c53cf7c2cda1efa8412458c4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebdc21df16a83d6df0f4e21a", "attempt_id": "746b7eafe3da", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:23.239672Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 524, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 364.4478749483824, "ttft_ms": 36.631166003644466, "decode_ms": 24.296708987094462}, "total_wall_ms": 87338.5049579665, "ttft_ms": 36.631166003644466, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3ced1f18c12704b1cf024101", "attempt_id": "c44c4249e85e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:23.684967Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 522, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 320.32225001603365, "ttft_ms": 36.599415936507285, "decode_ms": 58.985917014069855}, "total_wall_ms": 87329.00875003543, "ttft_ms": 36.599415936507285, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bef70d40e293f4fe6cba2d5a", "attempt_id": "38b23919d5fd", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:24.120586Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 538, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 492.3237500479445, "ttft_ms": 41.62766691297293, "decode_ms": 35.9469159739092}, "total_wall_ms": 87482.92862507515, "ttft_ms": 41.62766691297293, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "88d85891a6de24087ab8d04f", "attempt_id": "75dc3c5a34b0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:24.710996Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 530, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 352.77366696391255, "ttft_ms": 25.617250008508563, "decode_ms": 35.9563329257071}, "total_wall_ms": 87327.41116697434, "ttft_ms": 25.617250008508563, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "325fb786b9df2ea49220daf9", "attempt_id": "3e61faf995b1", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:25.145240Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 539, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 349.9067920492962, "ttft_ms": 19.871999975293875, "decode_ms": 44.94691709987819}, "total_wall_ms": 87327.82850007061, "ttft_ms": 19.871999975293875, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3a126fa10008a270caa32877", "attempt_id": "cf4b35b1a251", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:25.579418Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 534, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 360.09529104921967, "ttft_ms": 36.544166039675474, "decode_ms": 513.5753329377621}, "total_wall_ms": 87823.28045798931, "ttft_ms": 36.544166039675474, "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": "13,758,827", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "21aa0f915775b496e1fdf23385b8745829399f4cca3ef8a4e7b034514b836de7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "07f0945c5637d184df6aba67", "attempt_id": "5793f92b7919", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:26.510892Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 529, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 371.17004103492945, "ttft_ms": 25.189416017383337, "decode_ms": 120.26308302301913}, "total_wall_ms": 87429.85033406876, "ttft_ms": 25.189416017383337, "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": "7640", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b7aa5cfd1d2529920befa138d570102a53952edfdd4823a958212feea0698adf"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2e004fb175588af9fb212925", "attempt_id": "19064126b146", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:27.046919Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 534, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 360.8928339090198, "ttft_ms": 57.67258396372199, "decode_ms": 58.587040985003114}, "total_wall_ms": 87390.35387511831, "ttft_ms": 57.67258396372199, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "326ecf2666aa0e347482ac46", "attempt_id": "c583bf24caa0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:27.545679Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 535, "message_tokens": 440, "message_utf8_bytes": 983, "cache_tensor_bytes": null, "serialized_payload_bytes": 983, "side_information_bytes": 0, "fetched_bytes": 2367, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 86912.37875004299, "hops": {"1": 23978.786875028163, "2": 21622.011792031117, "3": 21589.40137492027, "4": 19722.17870806344}, "receiver_prefill_ms": 487.0039999950677, "ttft_ms": 34.92179198656231, "decode_ms": 46.30433302372694}, "total_wall_ms": 87481.25170799904, "ttft_ms": 34.92179198656231, "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": {"ledger": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 28, "retention_rate": 0.4375, "retention_by_kind": {"person": 0.6666666666666666, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1196}, "2": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1076}, "3": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1056}, "4": {"facts_total": 64, "facts_retained": 27, "retention_rate": 0.421875, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.42857142857142855, "status": 0.75, "place": 0.0, "identifier": 0.5, "quantity": 0.5}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 983}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "f3f9100f499c99c455c1d7e20c00120d9829a9cdac2bf0da8d125f5cea5b5e18", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "654738dd738843b0f907b0fe", "attempt_id": "c59d7d2e9499", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:51.920431Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 422, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 301.7415420617908, "ttft_ms": 40.49966693855822, "decode_ms": 26.65550005622208}, "total_wall_ms": 24175.431499956176, "ttft_ms": 40.49966693855822, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "86083f107a830c56fbbaf034", "attempt_id": "4b083d2b6999", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:52.312095Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 428, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 272.64125004876405, "ttft_ms": 44.94387493468821, "decode_ms": 59.50470897369087}, "total_wall_ms": 24183.52562491782, "ttft_ms": 44.94387493468821, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5e4ed0fac7baa9c78cd18d1a", "attempt_id": "62d2206643ee", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:52.707771Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 428, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 412.9656250588596, "ttft_ms": 73.26374994590878, "decode_ms": 24.701875052414834}, "total_wall_ms": 24317.420208943076, "ttft_ms": 73.26374994590878, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4113fc21d6235061ed56e51a", "attempt_id": "7862d413bb91", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:53.237999Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 425, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 289.05470902100205, "ttft_ms": 33.48912496585399, "decode_ms": 34.545957925729454}, "total_wall_ms": 24163.412792026065, "ttft_ms": 33.48912496585399, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "65982a3717fca2a3f6362e7d", "attempt_id": "c2c97d0d3b13", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:53.613889Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 433, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 289.6359999431297, "ttft_ms": 36.93308401852846, "decode_ms": 253.97008296567947}, "total_wall_ms": 24387.32970901765, "ttft_ms": 36.93308401852846, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a0319318d327a2dda1eee314", "attempt_id": "9f2d5b2ab129", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:54.212695Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 361.2338330131024, "ttft_ms": 135.44333295430988, "decode_ms": 343.8116249162704}, "total_wall_ms": 24646.979333017953, "ttft_ms": 135.44333295430988, "input_usage": {"lookups": 0}, "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": "BH-71Y-2851", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "38b55d8642902ac72f5987064e67fbe9b766a04dc338df11f6fa45d4cb64179e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd6fbf74c496e288ac0d1e8b", "attempt_id": "9275eb78ca8a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:55.072157Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 425, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 289.44566706195474, "ttft_ms": 36.93337505683303, "decode_ms": 26.956874993629754}, "total_wall_ms": 24159.747208002955, "ttft_ms": 36.93337505683303, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e7f92d2b09caeb7eaf57f837", "attempt_id": "415af800775e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:55.444327Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 423, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 277.7679580030963, "ttft_ms": 40.73979100212455, "decode_ms": 63.351582968607545}, "total_wall_ms": 24188.366541988216, "ttft_ms": 40.73979100212455, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3fb235fa3a0e8e45ef52fbb3", "attempt_id": "ddc2c8cf6529", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:55.845130Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 439, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 424.406792037189, "ttft_ms": 34.30800000205636, "decode_ms": 39.54937506932765}, "total_wall_ms": 24304.763499996625, "ttft_ms": 34.30800000205636, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3d25747adc4c9cde3d658e00", "attempt_id": "fcfc5f2c20ba", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:56.363799Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 286.45245800726116, "ttft_ms": 38.3228330174461, "decode_ms": 38.48087496589869}, "total_wall_ms": 24169.76999992039, "ttft_ms": 38.3228330174461, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bb5f615c98213de643746652", "attempt_id": "04f0cb379b0d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:56.748212Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 280.1847499795258, "ttft_ms": 22.066625067964196, "decode_ms": 34.62270798627287}, "total_wall_ms": 24143.589708954096, "ttft_ms": 22.066625067964196, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a40f39945f5a4625ff61207f", "attempt_id": "0fa49140f42c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:57.106473Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 435, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 307.44291690643877, "ttft_ms": 55.12120807543397, "decode_ms": 431.2700000591576}, "total_wall_ms": 24600.48495803494, "ttft_ms": 55.12120807543397, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "da28eaa07e4907ae759cd2c8", "attempt_id": "6654aa117e90", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:57.921897Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 430, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 283.76887494232506, "ttft_ms": 37.946667056530714, "decode_ms": 30.016290955245495}, "total_wall_ms": 24158.164208056405, "ttft_ms": 37.946667056530714, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "29fd0d9e991b2bb95ff3d22a", "attempt_id": "82d6b81926b7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:58.293976Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 435, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 286.6809159750119, "ttft_ms": 36.97725001256913, "decode_ms": 25.88037494570017}, "total_wall_ms": 24156.047665979713, "ttft_ms": 36.97725001256913, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fb70fa618fd955cb1271fbc9", "attempt_id": "b9a3b715cc2d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:09:58.662764Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 436, "message_tokens": 340, "message_utf8_bytes": 820, "cache_tensor_bytes": null, "serialized_payload_bytes": 820, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23805.876999977045, "hops": {"1": 23805.876999977045}, "receiver_prefill_ms": 282.15804195497185, "ttft_ms": 52.90920799598098, "decode_ms": 51.37399991508573}, "total_wall_ms": 24192.923499969766, "ttft_ms": 52.90920799598098, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "14d044fe2f36e98bd83ebd20864056431cd0a2879251cf98f1050dc7ef042296", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d2777274844a91b68b157955", "attempt_id": "94db2ca20bd5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:15.080307Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 279.47962505277246, "ttft_ms": 39.09325005952269, "decode_ms": 33.138667000457644}, "total_wall_ms": 40172.51520801801, "ttft_ms": 39.09325005952269, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ea62921c5b382def0e9c4175", "attempt_id": "49d844dc5688", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:15.453021Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 437, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 283.94999995362014, "ttft_ms": 40.39512504823506, "decode_ms": 36.43700003158301}, "total_wall_ms": 40181.56637600623, "ttft_ms": 40.39512504823506, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a69d9673243cc7c76db396d", "attempt_id": "cf9f477d8322", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:15.832150Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 437, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 277.9293330386281, "ttft_ms": 65.70133403874934, "decode_ms": 61.300290981307626}, "total_wall_ms": 40225.920208962634, "ttft_ms": 65.70133403874934, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a0aed64014ceb7ccef553c3c", "attempt_id": "4e4a1064e988", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:16.256960Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 434, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 443.508040974848, "ttft_ms": 21.26141602639109, "decode_ms": 38.26458298135549}, "total_wall_ms": 40324.05845797621, "ttft_ms": 21.26141602639109, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "500f807f73a4573e1ae966bf", "attempt_id": "3dfc2795799f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:16.781921Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 442, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 301.57045903615654, "ttft_ms": 38.95058308262378, "decode_ms": 236.21420795097947}, "total_wall_ms": 40397.63833396137, "ttft_ms": 38.95058308262378, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "04a72f29be9614d829084f06", "attempt_id": "0042e8750659", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:17.378019Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 302.858000039123, "ttft_ms": 54.75075007416308, "decode_ms": 55.15749996993691}, "total_wall_ms": 40233.601166983135, "ttft_ms": 54.75075007416308, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "46b47c73a8ec08698e26dddb", "attempt_id": "c064f9d2cc2e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:17.809953Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 434, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 425.35124998539686, "ttft_ms": 23.497166926972568, "decode_ms": 37.1018749428913}, "total_wall_ms": 40306.79066688754, "ttft_ms": 23.497166926972568, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "287843430e19c61699447a55", "attempt_id": "485d679815e7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:18.318583Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 432, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 286.06570803094655, "ttft_ms": 41.1287909373641, "decode_ms": 31.177625060081482}, "total_wall_ms": 40179.136833990924, "ttft_ms": 41.1287909373641, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ae79836405ec08626fa7fcf5", "attempt_id": "7dbd491d35fa", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:18.695823Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 448, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 283.1887500360608, "ttft_ms": 30.179332941770554, "decode_ms": 19.644124899059534}, "total_wall_ms": 40153.899124939926, "ttft_ms": 30.179332941770554, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4e16e91b72ba67b28f912e78", "attempt_id": "d28884ebdafc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:19.048520Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 271.7770410235971, "ttft_ms": 56.85787496622652, "decode_ms": 66.04920804966241}, "total_wall_ms": 40215.564416954294, "ttft_ms": 56.85787496622652, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "be35fb6250d4e4d9a0e2a9e8", "attempt_id": "ce1aeac9f0b2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:19.465653Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 449, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 419.7006250033155, "ttft_ms": 41.00787499919534, "decode_ms": 33.83095795288682}, "total_wall_ms": 40315.511791966856, "ttft_ms": 41.00787499919534, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "78ccd67707f4299a2fccb272", "attempt_id": "efe1b92682d4", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:19.981332Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 444, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 272.03637501224875, "ttft_ms": 39.24974997062236, "decode_ms": 222.08004200365394}, "total_wall_ms": 40354.416791931726, "ttft_ms": 39.24974997062236, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c1423501100f9bc34cd32a2d", "attempt_id": "250b09ac05d0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:20.536419Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 439, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 310.4549159761518, "ttft_ms": 50.19120802171528, "decode_ms": 63.30512498971075}, "total_wall_ms": 40244.94312494062, "ttft_ms": 50.19120802171528, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ba457c77ca69356bd00b993a", "attempt_id": "e7fe28fcd071", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:20.983980Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 444, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 429.67774998396635, "ttft_ms": 37.82308392692357, "decode_ms": 36.893125041387975}, "total_wall_ms": 40325.321916956455, "ttft_ms": 37.82308392692357, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cadf84fb3c5a2c05c97bbe14", "attempt_id": "474045511ea5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:21.512804Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 445, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 540, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39820.30066696461, "hops": {"1": 23805.876999977045, "2": 16014.423666987568}, "receiver_prefill_ms": 276.4661250403151, "ttft_ms": 39.44229206535965, "decode_ms": 38.384499959647655}, "total_wall_ms": 40175.1867919229, "ttft_ms": 39.44229206535965, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6eb861ab7ea020c83913e4b4", "attempt_id": "8a05a9f62b48", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:53.637147Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 431, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 366.6160419816151, "ttft_ms": 79.31408297736198, "decode_ms": 81.99395891278982}, "total_wall_ms": 72099.09916704055, "ttft_ms": 79.31408297736198, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5b4a1bbe298c7fcc85236bcc", "attempt_id": "08b7ef064d44", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:54.186455Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 437, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 271.4111249661073, "ttft_ms": 35.93254100997001, "decode_ms": 24.396082968451083}, "total_wall_ms": 71902.87279197946, "ttft_ms": 35.93254100997001, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eaf3010d488f76cfed6f4d83", "attempt_id": "5091bf284e97", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:54.539967Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 437, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 277.84379199147224, "ttft_ms": 21.31333400029689, "decode_ms": 38.686207961291075}, "total_wall_ms": 71909.07662489917, "ttft_ms": 21.31333400029689, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d75bd722e477deef2183679f", "attempt_id": "73f4513627bb", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:54.898133Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 434, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 258.20329203270376, "ttft_ms": 39.4431670429185, "decode_ms": 36.33358399383724}, "total_wall_ms": 71905.29025101569, "ttft_ms": 39.4431670429185, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2dd2cf3f9ddf276a5ef1fe52", "attempt_id": "d22612daf967", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:55.255167Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 442, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 305.45304098632187, "ttft_ms": 95.7969999872148, "decode_ms": 336.7052080575377}, "total_wall_ms": 72309.25733293407, "ttft_ms": 95.7969999872148, "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": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c981623ebe679daa56eaa0cc", "attempt_id": "0eb13c2d3574", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:56.014599Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 299.9752080067992, "ttft_ms": 25.71820910088718, "decode_ms": 355.3281669737771}, "total_wall_ms": 72252.48766690493, "ttft_ms": 25.71820910088718, "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": "BH-71Y-2851", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "38b55d8642902ac72f5987064e67fbe9b766a04dc338df11f6fa45d4cb64179e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "08e5178e652ad4ae98208917", "attempt_id": "1f014e8ce276", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:56.715650Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 434, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 276.38300007674843, "ttft_ms": 106.55733395833522, "decode_ms": 80.80900006461889}, "total_wall_ms": 72034.90650001913, "ttft_ms": 106.55733395833522, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "939bd03a7f4692c101c43412", "attempt_id": "cf33c3c62003", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:57.201782Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 432, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 343.36175001226366, "ttft_ms": 32.83545910380781, "decode_ms": 35.51412501838058}, "total_wall_ms": 71982.86504193675, "ttft_ms": 32.83545910380781, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cf708e9d4bdd03e701f78663", "attempt_id": "8e1dffb1dfa2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:57.633912Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 448, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 284.2248330125585, "ttft_ms": 26.247833971865475, "decode_ms": 37.79537498485297}, "total_wall_ms": 71919.31962489616, "ttft_ms": 26.247833971865475, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a3510aea4861be3503e21411", "attempt_id": "d8cbdb5d4122", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:58.003236Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 440, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 261.17262500338256, "ttft_ms": 55.64108397811651, "decode_ms": 55.07308407686651}, "total_wall_ms": 71943.10462498106, "ttft_ms": 55.64108397811651, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "40b6116cf0e52f0bcf450648", "attempt_id": "671fe94801db", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:58.394436Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 449, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 351.23037500306964, "ttft_ms": 132.48508295509964, "decode_ms": 39.21070892829448}, "total_wall_ms": 72094.11354199983, "ttft_ms": 132.48508295509964, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2b30b83865accdf69be12cb9", "attempt_id": "ba2ce3634ce2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:58.938387Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 444, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 288.17058296408504, "ttft_ms": 29.614917002618313, "decode_ms": 266.9886670773849}, "total_wall_ms": 72155.88274993934, "ttft_ms": 29.614917002618313, "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": "7280585", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "67860e1409de7b24ac27a0ededca1d8f10c1a6a620ff6743e01199e2fb986943"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6b7b9171448bbbbabb2aaf2e", "attempt_id": "098dab892a84", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:59.542613Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 439, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 271.4880419662222, "ttft_ms": 65.12233300600201, "decode_ms": 59.99966606032103}, "total_wall_ms": 71967.87679195404, "ttft_ms": 65.12233300600201, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fbcb97af0dfdb1750b79ac50", "attempt_id": "ca917de5d207", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:10:59.959330Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 444, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 319.9792919913307, "ttft_ms": 136.74441596958786, "decode_ms": 45.7414579577744}, "total_wall_ms": 72073.68199992925, "ttft_ms": 136.74441596958786, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3d363ba79c7aece891923dd8", "attempt_id": "9c70ee5824cc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:00.483401Z", "task_family": "synthetic", "document_id": "syn_02000_2000", "document_hash": "3bcf1bb02792129efca162575bc2ef64622b41083d4e2870fed106d663e6417c", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02000_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1971, "query_tokens": 445, "message_tokens": 350, "message_utf8_bytes": 849, "cache_tensor_bytes": null, "serialized_payload_bytes": 849, "side_information_bytes": 0, "fetched_bytes": 1620, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 71570.5262919655, "hops": {"1": 23805.876999977045, "2": 16014.423666987568, "3": 15973.476000013761, "4": 15776.749624987133}, "receiver_prefill_ms": 281.9139170460403, "ttft_ms": 19.628999987617135, "decode_ms": 40.682208025828004}, "total_wall_ms": 71913.32608391531, "ttft_ms": 19.628999987617135, "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": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 820}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.25, "amount": 0.07142857142857142, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 849}}}, "reference_key_id": "keys/synthetic/syn_02000_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/E4_pilot.log", "payload_hash": "b43c43ee90d63cc5e7f6b246557f4b1068ae129211609ebdd647596a60b8fce5", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "11f8030c08ab16c9a1a96917", "attempt_id": "6a460e2e2f9b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:27.476098Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 420.57525005657226, "ttft_ms": 43.65054203663021, "decode_ms": 432.9306250438094}, "total_wall_ms": 27360.997540992685, "ttft_ms": 43.65054203663021, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0cd6d51c241581d9a8fd3b64", "attempt_id": "fa58a8b73474", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:28.400533Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 410.0057090399787, "ttft_ms": 38.426625076681376, "decode_ms": 30.01241700258106}, "total_wall_ms": 26942.012208048254, "ttft_ms": 38.426625076681376, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "936a01e73d72c6587dc83700", "attempt_id": "0a15dcba9da7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:28.903325Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 356.9008329650387, "ttft_ms": 52.668999996967614, "decode_ms": 44.859500019811094}, "total_wall_ms": 26918.04612497799, "ttft_ms": 52.668999996967614, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3702567f0f9150e0d800570e", "attempt_id": "ccf518de7c8c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:29.381338Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 534.7057910403237, "ttft_ms": 40.19800003152341, "decode_ms": 176.54620890971273}, "total_wall_ms": 27215.15641605947, "ttft_ms": 40.19800003152341, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1e5811e56d7686ee016b377a", "attempt_id": "245a30b8f628", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:30.158268Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 398.1017080368474, "ttft_ms": 19.222875009290874, "decode_ms": 41.240792023018}, "total_wall_ms": 26922.105040983297, "ttft_ms": 19.222875009290874, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "11b9226ae95d7f67b4e5a41d", "attempt_id": "1f135d3e3e6a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:30.640239Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 364.3018329748884, "ttft_ms": 43.47858391702175, "decode_ms": 475.5115000298247}, "total_wall_ms": 27346.985165961087, "ttft_ms": 43.47858391702175, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "77eda9bc01558899efd26239", "attempt_id": "12a38deca7c8", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:31.548636Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 386.1028749961406, "ttft_ms": 31.691250042058527, "decode_ms": 35.00879101920873}, "total_wall_ms": 26916.369624901563, "ttft_ms": 31.691250042058527, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e89ae234fe6e8f3616469cca", "attempt_id": "80b00a2f93a3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:32.026155Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 364.13083400111645, "ttft_ms": 21.45491703413427, "decode_ms": 66.34925003163517}, "total_wall_ms": 26915.517707937397, "ttft_ms": 21.45491703413427, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8b2c9dbe8fbfbe09d789cdc2", "attempt_id": "752e582d04b2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:32.502313Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 381.3158329576254, "ttft_ms": 91.77616599481553, "decode_ms": 317.79654195997864}, "total_wall_ms": 27254.67758299783, "ttft_ms": 91.77616599481553, "input_usage": {"lookups": 0}, "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": "755615", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "8e5f423d26733b737903e04c5953aa05727c21f93ccd85f8b611a4116e99673e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "47280f033b652272b7723d74", "attempt_id": "889adb235078", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:33.318469Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 393.9206668874249, "ttft_ms": 35.67762498278171, "decode_ms": 224.78445910383016}, "total_wall_ms": 27118.014916894026, "ttft_ms": 35.67762498278171, "input_usage": {"lookups": 0}, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6ad9d59908a2addaa41379ae", "attempt_id": "ee0420e5ad6c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:33.999828Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 368.02279099356383, "ttft_ms": 37.8275829134509, "decode_ms": 59.96366695035249}, "total_wall_ms": 26929.52904093545, "ttft_ms": 37.8275829134509, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ad6c521c886b2c41e32ae9b3", "attempt_id": "5f8014f763c3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:34.492940Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 519.3456249544397, "ttft_ms": 24.07987497281283, "decode_ms": 214.30666698142886}, "total_wall_ms": 27221.417582943104, "ttft_ms": 24.07987497281283, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8d41cf7534ac36a28a728cd4", "attempt_id": "3931d701655a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:35.277518Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 377.5448329979554, "ttft_ms": 22.981708985753357, "decode_ms": 316.65024999529123}, "total_wall_ms": 27180.885375011712, "ttft_ms": 22.981708985753357, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "253c5303c845943b7535e09c", "attempt_id": "3db1456fdabf", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:36.019938Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 505.8825839078054, "ttft_ms": 39.34937505982816, "decode_ms": 23.443207959644496}, "total_wall_ms": 27032.321415957995, "ttft_ms": 39.34937505982816, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "89c92a276078d8699f851f0a", "attempt_id": "b0a855aa5c49", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:36.612767Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26462.733207968995, "hops": {"1": 26462.733207968995}, "receiver_prefill_ms": 386.4243330899626, "ttft_ms": 37.354541942477226, "decode_ms": 26.947916019707918}, "total_wall_ms": 26914.33145792689, "ttft_ms": 37.354541942477226, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "12a46507e7dc3c55deb57652", "attempt_id": "b01e1b063b18", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:56.366132Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 511, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 324.19600000139326, "ttft_ms": 38.744290941394866, "decode_ms": 289.8643750231713}, "total_wall_ms": 46393.09204183519, "ttft_ms": 38.744290941394866, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "20f8dbdb6f7b194563d5d7a4", "attempt_id": "3ebc39387c28", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:57.044062Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 516.5617089951411, "ttft_ms": 36.19070793502033, "decode_ms": 21.376416087150574}, "total_wall_ms": 46314.353666850366, "ttft_ms": 36.19070793502033, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1b5ac50c3b935f5534d7bed1", "attempt_id": "17f532c6a4b4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:57.642609Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 517, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 362.47120797634125, "ttft_ms": 40.871959063224494, "decode_ms": 30.503207934089005}, "total_wall_ms": 46174.1996249184, "ttft_ms": 40.871959063224494, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "662523aa70cbe3154f5a7928", "attempt_id": "11c3ef79c25d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:58.101340Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 512, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 302.08262510132045, "ttft_ms": 21.61320799496025, "decode_ms": 249.5709580834955}, "total_wall_ms": 46313.710624934174, "ttft_ms": 21.61320799496025, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e28ab76faa901e6c6128c025", "attempt_id": "c1e2b7410314", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:58.700102Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 496.3327079312876, "ttft_ms": 29.236541013233364, "decode_ms": 41.499959072098136}, "total_wall_ms": 46307.418583892286, "ttft_ms": 29.236541013233364, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6fb3f870309a6fec3256f992", "attempt_id": "0216a489a229", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:59.292182Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 521, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 371.1953329620883, "ttft_ms": 25.079249986447394, "decode_ms": 228.06083306204528}, "total_wall_ms": 46364.459250937216, "ttft_ms": 25.079249986447394, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "17504e1587cbf3960e1fd96e", "attempt_id": "52c68bc488f3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:11:59.941032Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 512, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 313.8362920144573, "ttft_ms": 55.76479190494865, "decode_ms": 60.568541986867785}, "total_wall_ms": 46170.5006249249, "ttft_ms": 55.76479190494865, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "334875d6f7ffc92139c05ec2", "attempt_id": "40e545832ead", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:00.396683Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 514, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 497.2001670394093, "ttft_ms": 39.93854089640081, "decode_ms": 31.18050005286932}, "total_wall_ms": 46308.56029188726, "ttft_ms": 39.93854089640081, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "19a7749b449f50d950e4b341", "attempt_id": "3f22c15a33c2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:00.989715Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 529, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 368.5186670627445, "ttft_ms": 37.697458057664335, "decode_ms": 38.24741707649082}, "total_wall_ms": 46184.90574986208, "ttft_ms": 37.697458057664335, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2264bfde2ed65b9fb7518ecc", "attempt_id": "e6ff53f26c20", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:01.459004Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 523, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 338.0878750467673, "ttft_ms": 64.94283396750689, "decode_ms": 489.27349993027747}, "total_wall_ms": 46632.56350089796, "ttft_ms": 64.94283396750689, "input_usage": {"lookups": 0}, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7d24666d1f67fabfb74a6be6", "attempt_id": "1bc29e05dcd7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:02.379354Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 530, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 362.6911669271067, "ttft_ms": 31.907832948490977, "decode_ms": 36.89545893575996}, "total_wall_ms": 46171.658249921165, "ttft_ms": 31.907832948490977, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3b66744c28b391173b186d93", "attempt_id": "bfae70b70b89", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:02.834458Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 524, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 345.17575008794665, "ttft_ms": 15.415457892231643, "decode_ms": 262.9735419759527}, "total_wall_ms": 46363.80637483671, "ttft_ms": 15.415457892231643, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "234c2f15299ef78816ddc6d8", "attempt_id": "cee981808ed2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:03.482894Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 512.5960829900578, "ttft_ms": 37.790499976836145, "decode_ms": 190.83766592666507}, "total_wall_ms": 46481.393624912016, "ttft_ms": 37.790499976836145, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a4d809e2d3fb8ca0b7f85dab", "attempt_id": "8894f3f07e06", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:04.248727Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 523, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 351.23266698792577, "ttft_ms": 31.936417100951076, "decode_ms": 35.45233397744596}, "total_wall_ms": 46158.87087583542, "ttft_ms": 31.936417100951076, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e935fb48288e67312bede2fd", "attempt_id": "e02d2ba1b2f4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:04.691491Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 526, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45739.48579188436, "hops": {"1": 26462.733207968995, "2": 19276.752583915368}, "receiver_prefill_ms": 367.08579189144075, "ttft_ms": 55.51700002979487, "decode_ms": 55.54962495807558}, "total_wall_ms": 46218.503791838884, "ttft_ms": 55.51700002979487, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "30ec81041783499852221ae1", "attempt_id": "702efdcf40a0", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:43.353240Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 511, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 313.9787499094382, "ttft_ms": 39.12387497257441, "decode_ms": 40.43000005185604}, "total_wall_ms": 84287.8709579818, "ttft_ms": 39.12387497257441, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "234c368d49acf5ddddadbf91", "attempt_id": "7bb6b2b04762", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:43.770680Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 354.43587496411055, "ttft_ms": 54.14712498895824, "decode_ms": 58.34304098971188}, "total_wall_ms": 84361.42533295788, "ttft_ms": 54.14712498895824, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d9929aa8eb6a3c463cac3294", "attempt_id": "37e7712e3aa7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:44.262545Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 517, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 418.40924997814, "ttft_ms": 112.29454097338021, "decode_ms": 32.404916011728346}, "total_wall_ms": 84457.51999889035, "ttft_ms": 112.29454097338021, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8104dcbc3c5025d1431c121c", "attempt_id": "4cc6fe1423fa", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:44.849560Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 512, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 296.5612079715356, "ttft_ms": 37.915208959020674, "decode_ms": 161.5741669666022}, "total_wall_ms": 84390.51537390333, "ttft_ms": 37.915208959020674, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1b94df49ba07fd5ebe1b5588", "attempt_id": "2e9e58134fb3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:45.369654Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 337.6168749528006, "ttft_ms": 51.49008391890675, "decode_ms": 55.50387501716614}, "total_wall_ms": 84339.06083286274, "ttft_ms": 51.49008391890675, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ec6fc6a617498ba754756c78", "attempt_id": "889facb73044", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:45.838523Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 521, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 387.4572080094367, "ttft_ms": 90.25949996430427, "decode_ms": 303.86054201517254}, "total_wall_ms": 84675.93154089991, "ttft_ms": 90.25949996430427, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b30fd8e6e71055f29e9eb11c", "attempt_id": "a3b4ed0bdf58", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:46.644292Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 512, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 298.2711670920253, "ttft_ms": 38.556624902412295, "decode_ms": 24.635499925352633}, "total_wall_ms": 84255.82537497394, "ttft_ms": 38.556624902412295, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d99750240eff59493a2fc49e", "attempt_id": "43e1da8180e6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:47.029354Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 514, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 336.95304091088474, "ttft_ms": 59.1314589837566, "decode_ms": 51.68012506328523}, "total_wall_ms": 84342.11104095448, "ttft_ms": 59.1314589837566, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "249b48112580b8b8b282b3e0", "attempt_id": "04e81c5c55e4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:47.500250Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 529, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 440.677999984473, "ttft_ms": 94.87579204142094, "decode_ms": 36.160957999527454}, "total_wall_ms": 84466.02841594722, "ttft_ms": 94.87579204142094, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "50f69fe2fe87becba0f43ccc", "attempt_id": "3377e45b62e8", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:48.094908Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 523, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 352.7269170153886, "ttft_ms": 36.310457973740995, "decode_ms": 290.0730420369655}, "total_wall_ms": 84573.51316593122, "ttft_ms": 36.310457973740995, "input_usage": {"lookups": 0}, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c69ac045c5e04e2a2421f240", "attempt_id": "ac3d5520c050", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:48.798348Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 530, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 316.3985839346424, "ttft_ms": 35.691707977093756, "decode_ms": 55.05966593045741}, "total_wall_ms": 84301.52895790525, "ttft_ms": 35.691707977093756, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb20ffe410c9e0f11cc9e79f", "attempt_id": "88b0c534e8c2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:49.229619Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 524, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 500.88795798365027, "ttft_ms": 26.06370800640434, "decode_ms": 216.5380419464782}, "total_wall_ms": 84637.93116586749, "ttft_ms": 26.06370800640434, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "13104d0ce8eb73d42b14dff3", "attempt_id": "24bd502c9711", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:49.997318Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 378.3195000141859, "ttft_ms": 37.04183304216713, "decode_ms": 181.50216701906174}, "total_wall_ms": 84491.27720692195, "ttft_ms": 37.04183304216713, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c39f32caf84c16d1ec7dc3d3", "attempt_id": "cf6b31e8046b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:50.618810Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 523, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 382.2503339033574, "ttft_ms": 86.92708297166973, "decode_ms": 93.11945806257427}, "total_wall_ms": 84456.84716594405, "ttft_ms": 86.92708297166973, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3fe5f68c5abf59f8e2453fe8", "attempt_id": "daa762e53626", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:12:51.206452Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 526, "message_tokens": 430, "message_utf8_bytes": 1461, "cache_tensor_bytes": null, "serialized_payload_bytes": 1461, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 83893.65316589829, "hops": {"1": 26462.733207968995, "2": 19276.752583915368, "3": 18933.089833008125, "4": 19221.077541005798}, "receiver_prefill_ms": 354.6070419251919, "ttft_ms": 40.228125057183206, "decode_ms": 30.713167041540146}, "total_wall_ms": 84320.08441584185, "ttft_ms": 40.228125057183206, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "3": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}, "4": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1461}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "2b700983f8d010251bb6fe845c89b8acbfece8a8ae1264729ea857d72d728284", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "439bc80b1fede65003bc7256", "attempt_id": "fb25eb2f975f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:18.142873Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 259.3166249571368, "ttft_ms": 52.73608292918652, "decode_ms": 41.685375035740435}, "total_wall_ms": 26836.69899997767, "ttft_ms": 52.73608292918652, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7fbca90e8b549549e5c78480", "attempt_id": "0ddcd9d330a3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:18.522585Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 361.99837492313236, "ttft_ms": 94.59212492220104, "decode_ms": 88.4642080636695}, "total_wall_ms": 27027.627875097096, "ttft_ms": 94.59212492220104, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "47885edce5298abf39538069", "attempt_id": "866f1e778b02", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:19.094470Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 441, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 282.0017089834437, "ttft_ms": 38.942792103625834, "decode_ms": 40.39300000295043}, "total_wall_ms": 26843.985832994804, "ttft_ms": 38.942792103625834, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "711a0790e1ea0ccffe7211e3", "attempt_id": "65950cc2eab9", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:19.480063Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 283.39195798616856, "ttft_ms": 33.850249950774014, "decode_ms": 139.04691708739847}, "total_wall_ms": 26938.90133302193, "ttft_ms": 33.850249950774014, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1f450d80678073d77be5e9a0", "attempt_id": "75cc3bbfaeba", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:19.961082Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 316.88166595995426, "ttft_ms": 55.808958946727216, "decode_ms": 110.83474999759346}, "total_wall_ms": 26966.154167079367, "ttft_ms": 55.808958946727216, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "70953a3e20501444db8341c6", "attempt_id": "a24cae51255b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:20.470281Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 445, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 362.42179095279425, "ttft_ms": 38.33775001112372, "decode_ms": 37.42229100316763}, "total_wall_ms": 26920.65891600214, "ttft_ms": 38.33775001112372, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "50f9b15fc3fa1512510ac51f", "attempt_id": "d40f738ebe44", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:20.934606Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 297.31479100883007, "ttft_ms": 30.156166991218925, "decode_ms": 40.05033301655203}, "total_wall_ms": 26850.135750020854, "ttft_ms": 30.156166991218925, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7c6a2383bf9ea6ba1d41143d", "attempt_id": "96908c52826c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:21.325946Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 438, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 290.8776659751311, "ttft_ms": 35.13874998316169, "decode_ms": 54.22312498558313}, "total_wall_ms": 26862.821250106208, "ttft_ms": 35.13874998316169, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3dc0c7a489caf34c7b28f0de", "attempt_id": "e68436d74824", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:21.730154Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 453, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 355.55366694461554, "ttft_ms": 135.96241699997336, "decode_ms": 37.36708301585168}, "total_wall_ms": 27011.688290978782, "ttft_ms": 135.96241699997336, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bb0c840eb324179215ec7b77", "attempt_id": "c544f6dd7c1a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:22.283110Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 294.0003339899704, "ttft_ms": 41.028750012628734, "decode_ms": 25.806791963987052}, "total_wall_ms": 26843.33849896211, "ttft_ms": 41.028750012628734, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e34f3fc973603134b1fa89a3", "attempt_id": "bc8f556b8f5c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:22.667354Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 454, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 314.96458302717656, "ttft_ms": 34.77816702798009, "decode_ms": 29.507125029340386}, "total_wall_ms": 26861.942791030742, "ttft_ms": 34.77816702798009, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "15cada4051aa65650f4af9bd", "attempt_id": "e58412177d5e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:23.070316Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 448, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 267.0352919958532, "ttft_ms": 36.7047080071643, "decode_ms": 445.3237500274554}, "total_wall_ms": 27231.588416034356, "ttft_ms": 36.7047080071643, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "345ff387147bbc2bf4cc0de6", "attempt_id": "3eca03d5bfae", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:23.845111Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 276.7084169900045, "ttft_ms": 21.12737495917827, "decode_ms": 185.59579201973975}, "total_wall_ms": 26966.241499991156, "ttft_ms": 21.12737495917827, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "54679ae95251467b0fc655a5", "attempt_id": "bb40e0f1bcb1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:24.352257Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 272.34050002880394, "ttft_ms": 20.10337496176362, "decode_ms": 19.932541996240616}, "total_wall_ms": 26794.987665954977, "ttft_ms": 20.10337496176362, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3db38ec17270a97b6dca66cf", "attempt_id": "971e8fd83430", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:24.689728Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 450, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26481.940208002925, "hops": {"1": 26481.940208002925}, "receiver_prefill_ms": 284.5792080042884, "ttft_ms": 58.937499998137355, "decode_ms": 45.44095799792558}, "total_wall_ms": 26871.462957933545, "ttft_ms": 58.937499998137355, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f48a5ef9353beb19b91f6483", "attempt_id": "86777bf46a45", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:41.593768Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 257.1123749949038, "ttft_ms": 54.990667034871876, "decode_ms": 64.56833297852427}, "total_wall_ms": 43346.59695706796, "ttft_ms": 54.990667034871876, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "53fd9f50c70230c250ead9cb", "attempt_id": "bbd265d93c5f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:41.995457Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 470.6381659489125, "ttft_ms": 21.84512501116842, "decode_ms": 35.96808295696974}, "total_wall_ms": 43498.52062400896, "ttft_ms": 21.84512501116842, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3697fe4c2487525c6926424b", "attempt_id": "75cefaf6411f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:42.549729Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 441, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 286.6072079632431, "ttft_ms": 21.927875000983477, "decode_ms": 38.496125023812056}, "total_wall_ms": 43316.80716609117, "ttft_ms": 21.927875000983477, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2b581f3b13c981e6c4dd9935", "attempt_id": "8be4afd301b7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:42.920947Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 288.34374994039536, "ttft_ms": 21.502749994397163, "decode_ms": 198.11445893719792}, "total_wall_ms": 43477.65520703979, "ttft_ms": 21.502749994397163, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bb6a5d7cf50b9dbff799ecaf", "attempt_id": "fa1b61997f2c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:43.452564Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 276.9519160501659, "ttft_ms": 108.35337499156594, "decode_ms": 137.83074996899813}, "total_wall_ms": 43492.88624897599, "ttft_ms": 108.35337499156594, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0885b26df85a46b0b621bcf6", "attempt_id": "d2ea1528075f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:44.000507Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 445, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 275.67887492477894, "ttft_ms": 40.37454107310623, "decode_ms": 24.69383308198303}, "total_wall_ms": 43310.55504095275, "ttft_ms": 40.37454107310623, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "549aa03ee2ed5b1ddb1106bd", "attempt_id": "49664592c2fa", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:44.366049Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 277.3140409262851, "ttft_ms": 33.44025008846074, "decode_ms": 43.72741701081395}, "total_wall_ms": 43324.31166607421, "ttft_ms": 33.44025008846074, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "04d7e0a40d4f4d33d4c2049d", "attempt_id": "33df85e5ba15", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:44.744270Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 438, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 266.7712499387562, "ttft_ms": 63.66174994036555, "decode_ms": 48.96404198370874}, "total_wall_ms": 43349.17487401981, "ttft_ms": 63.66174994036555, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ffa502a651b12620b41975ca", "attempt_id": "8d5ed400664c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:45.148751Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 453, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 397.9043329600245, "ttft_ms": 98.25970896054059, "decode_ms": 21.585791022516787}, "total_wall_ms": 43487.51033202279, "ttft_ms": 98.25970896054059, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "250a4a0338cf4a2ee9b60864", "attempt_id": "c31d4ca2ae38", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:45.693133Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 284.10216700285673, "ttft_ms": 36.14104096777737, "decode_ms": 25.494041037745774}, "total_wall_ms": 43315.615499042906, "ttft_ms": 36.14104096777737, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "647aadf10c7e5bb14d762e17", "attempt_id": "3afd0952921a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:46.062274Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 454, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 306.8330839741975, "ttft_ms": 38.59283297788352, "decode_ms": 24.388333898968995}, "total_wall_ms": 43339.68158205971, "ttft_ms": 38.59283297788352, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3c36869f6e27225105976945", "attempt_id": "61bb5866f1ec", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:46.457312Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 448, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 257.571000023745, "ttft_ms": 56.100583053193986, "decode_ms": 442.7752500632778}, "total_wall_ms": 43726.13983193878, "ttft_ms": 56.100583053193986, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f8516c43b6a08e7b87fb2fd7", "attempt_id": "f4aa4736c1d3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:47.236394Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 293.3017920004204, "ttft_ms": 37.29991696309298, "decode_ms": 179.68895810190588}, "total_wall_ms": 43479.91695697419, "ttft_ms": 37.29991696309298, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "73e6e62c332c8b94fdd77f63", "attempt_id": "493203490de1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:47.770129Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 277.316416031681, "ttft_ms": 40.315540973097086, "decode_ms": 59.51987497974187}, "total_wall_ms": 43347.167456988245, "ttft_ms": 40.315540973097086, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8c3f91d49babdac81cfb72c1", "attempt_id": "d1df6dc11d43", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:13:48.171968Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 450, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42969.16137402877, "hops": {"1": 26481.940208002925, "2": 16487.221166025847}, "receiver_prefill_ms": 303.69145795702934, "ttft_ms": 93.84949994273484, "decode_ms": 87.92733401060104}, "total_wall_ms": 43455.2378320368, "ttft_ms": 93.84949994273484, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c44904fa8a89de04e60ed7fd", "attempt_id": "0c1c9459b139", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:21.440424Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 289.99362501781434, "ttft_ms": 37.51145803835243, "decode_ms": 38.45858294516802}, "total_wall_ms": 76089.62428988889, "ttft_ms": 37.51145803835243, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "711f6e8fbae32da7be9aeff4", "attempt_id": "70542d990633", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:21.830673Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 282.3493330506608, "ttft_ms": 40.761208976618946, "decode_ms": 23.457332979887724}, "total_wall_ms": 76070.12533198576, "ttft_ms": 40.761208976618946, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "787d42e8712e77903dc56efd", "attempt_id": "564a840524de", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:22.203936Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 441, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 280.63408308662474, "ttft_ms": 58.0794169800356, "decode_ms": 46.36100004427135}, "total_wall_ms": 76108.67099906318, "ttft_ms": 58.0794169800356, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "34ad7d0996290bf14a37c986", "attempt_id": "ded6e00aab1c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:22.614884Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 419.88358297385275, "ttft_ms": 32.49362495262176, "decode_ms": 160.62049998436123}, "total_wall_ms": 76336.67495695408, "ttft_ms": 32.49362495262176, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "20baeffdc814bf0e5851825f", "attempt_id": "e2bc3fd21a51", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:23.254465Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 275.51275002770126, "ttft_ms": 40.596332983113825, "decode_ms": 29.919540975242853}, "total_wall_ms": 76069.58774896339, "ttft_ms": 40.596332983113825, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "56ca40c928cb5a16901a19fe", "attempt_id": "29f960a41517", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:23.626487Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 445, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 245.83720904774964, "ttft_ms": 53.848791983909905, "decode_ms": 47.356957918964326}, "total_wall_ms": 76070.55191497784, "ttft_ms": 53.848791983909905, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cf979922e07a50b09ece2a4d", "attempt_id": "b78e44ca67fa", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:23.997295Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 276.4265830628574, "ttft_ms": 92.52129192464054, "decode_ms": 134.48224996682256}, "total_wall_ms": 76227.15670696925, "ttft_ms": 92.52129192464054, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ec4d97bb5fadbee5306b54dd", "attempt_id": "18ff21114a44", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:24.526259Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 438, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 273.23504199739546, "ttft_ms": 41.40208300668746, "decode_ms": 27.29099988937378}, "total_wall_ms": 76065.53395697847, "ttft_ms": 41.40208300668746, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "012ffde890fe74479d6a71fd", "attempt_id": "5e9da40b04e5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:24.891851Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 453, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 298.77133306581527, "ttft_ms": 32.20483404584229, "decode_ms": 34.14787503425032}, "total_wall_ms": 76088.69578992017, "ttft_ms": 32.20483404584229, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c6b92bafe851fcaf86c85106", "attempt_id": "f5e005f57b9b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:25.280479Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 266.93437492940575, "ttft_ms": 59.440000099129975, "decode_ms": 40.68041709251702}, "total_wall_ms": 76090.56333103217, "ttft_ms": 59.440000099129975, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dcea3db04f6ce6cbec033267", "attempt_id": "1c3a86ee76ae", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:25.672349Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 454, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 448.301958036609, "ttft_ms": 61.84945802669972, "decode_ms": 31.543999910354614}, "total_wall_ms": 76265.20766504109, "ttft_ms": 61.84945802669972, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e770e26daebcbaad1234e11c", "attempt_id": "8ecc714867a9", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:26.237955Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 448, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 282.06516604404896, "ttft_ms": 37.868292070925236, "decode_ms": 199.6369999833405}, "total_wall_ms": 76243.21128998417, "ttft_ms": 37.868292070925236, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7dc8b36617bfb3ff402f0a02", "attempt_id": "1df9699249e1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:26.780493Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 272.44537498336285, "ttft_ms": 39.81187508907169, "decode_ms": 285.09908297564834}, "total_wall_ms": 76320.93166490085, "ttft_ms": 39.81187508907169, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c2404fb28281c5a8cc97d9ca", "attempt_id": "f1a7d8a504c1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:27.401957Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 423.101625056006, "ttft_ms": 33.888707985170186, "decode_ms": 40.1537079596892}, "total_wall_ms": 76220.75762390159, "ttft_ms": 33.888707985170186, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eb94b5283c984b317e798025", "attempt_id": "4b78321340f0", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:27.923092Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 450, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 75722.94553997926, "hops": {"1": 26481.940208002925, "2": 16487.221166025847, "3": 16283.874499960802, "4": 16469.90966598969}, "receiver_prefill_ms": 319.33354202192277, "ttft_ms": 37.51616703812033, "decode_ms": 24.13787494879216}, "total_wall_ms": 76104.59866502788, "ttft_ms": 37.51616703812033, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "3": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "4": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ed23b5a9deb64c0961930ff5", "attempt_id": "1fc57b13ac0d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:55.351086Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 384.47087502572685, "ttft_ms": 43.2934999698773, "decode_ms": 235.8323340304196}, "total_wall_ms": 27685.226708068512, "ttft_ms": 43.2934999698773, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f01de663282d8d21bae58b07", "attempt_id": "62e94cbd16bc", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:56.039171Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 396.33925002999604, "ttft_ms": 53.93708299379796, "decode_ms": 45.56758294347674}, "total_wall_ms": 27517.46808399912, "ttft_ms": 53.93708299379796, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5e3956852dd8dc8899769cf7", "attempt_id": "87530696761b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:56.558356Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 542.8143750177696, "ttft_ms": 36.84154199436307, "decode_ms": 31.74070897512138}, "total_wall_ms": 27633.18229210563, "ttft_ms": 36.84154199436307, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a91f8dac417979c6bad01b8d", "attempt_id": "f04a32db818c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:57.194731Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 385.723749990575, "ttft_ms": 38.141541997902095, "decode_ms": 159.6637499751523}, "total_wall_ms": 27605.226375046186, "ttft_ms": 38.141541997902095, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "25c8da3cdab8c9b244826423", "attempt_id": "cba3f3796617", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:57.802263Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 371.7637090012431, "ttft_ms": 59.73275005817413, "decode_ms": 44.528374914079905}, "total_wall_ms": 27497.757375007495, "ttft_ms": 59.73275005817413, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "67124bf625b55cc5bb21ad83", "attempt_id": "404857503503", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:58.302572Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 529.6712080016732, "ttft_ms": 35.50550004001707, "decode_ms": 251.88837491441518}, "total_wall_ms": 27838.82695902139, "ttft_ms": 35.50550004001707, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "BJ-45Z-9", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "7c92b4ad493b083ec22ecc5c025a6f9e2648b8dfa0da12613f1648d3773a3306"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eb795e7f02ac75025bd39c96", "attempt_id": "243c5e6e8399", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:59.146671Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 380.60537504497916, "ttft_ms": 24.42470903042704, "decode_ms": 75.06979198660702}, "total_wall_ms": 27501.831082976423, "ttft_ms": 24.42470903042704, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "73e953cea71961c371460828", "attempt_id": "706a06abe9d0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:14:59.651456Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 368.96649992559105, "ttft_ms": 93.35904207546264, "decode_ms": 132.73937499616295}, "total_wall_ms": 27617.009958950803, "ttft_ms": 93.35904207546264, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d7ac95e0d0a806d419f07e96", "attempt_id": "1acd1bba199d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:00.273137Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 385.2736660046503, "ttft_ms": 41.93012497853488, "decode_ms": 239.38345804344863}, "total_wall_ms": 27688.29487496987, "ttft_ms": 41.93012497853488, "input_usage": {"lookups": 0}, "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": "755615", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "8e5f423d26733b737903e04c5953aa05727c21f93ccd85f8b611a4116e99673e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "013f09aea7c86c4798b185d2", "attempt_id": "b6592102aa5e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:00.966078Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 354.5337909599766, "ttft_ms": 43.95737498998642, "decode_ms": 528.2931659603491}, "total_wall_ms": 27948.478708043694, "ttft_ms": 43.95737498998642, "input_usage": {"lookups": 0}, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "acef178bc31c120f95bacc49", "attempt_id": "dc1670bde79d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:01.917636Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 394.34633299242705, "ttft_ms": 39.09004095476121, "decode_ms": 34.07037502620369}, "total_wall_ms": 27489.260792033747, "ttft_ms": 39.09004095476121, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ca3bca689c5c555ddb73d490", "attempt_id": "cdc35a10d945", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:02.412114Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 363.12545801047236, "ttft_ms": 61.40245799906552, "decode_ms": 284.1740410076454}, "total_wall_ms": 27730.62920803204, "ttft_ms": 61.40245799906552, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "161adde18418312d356d9976", "attempt_id": "f57ac3be525d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:03.144492Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 533.1462919712067, "ttft_ms": 40.964667103253305, "decode_ms": 192.7076670108363}, "total_wall_ms": 27788.697832962498, "ttft_ms": 40.964667103253305, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2eb08b1b4e0c4e1ef7c1fa67", "attempt_id": "e3ef55ac445b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:03.935075Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 377.7693329611793, "ttft_ms": 32.915249932557344, "decode_ms": 65.39766595233232}, "total_wall_ms": 27497.82904202584, "ttft_ms": 32.915249932557344, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d8cbb35e7880bc53f66cd2d9", "attempt_id": "bb1efa4ecbc2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:04.434889Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1813, "cache_tensor_bytes": null, "serialized_payload_bytes": 1813, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27020.851625013165, "hops": {"1": 27020.851625013165}, "receiver_prefill_ms": 477.1313340170309, "ttft_ms": 126.72854191623628, "decode_ms": 21.582042099907994}, "total_wall_ms": 27647.100166999735, "ttft_ms": 126.72854191623628, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "ee2605daa88ff1a6bb318582817ec771d595c06cee2eae85fde4f83c49fe232f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "493cab3cff266fb8fa20ab0b", "attempt_id": "e95e35459b8d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:28.130886Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 592, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 536.6073340410367, "ttft_ms": 33.20241696201265, "decode_ms": 36.79729194846004}, "total_wall_ms": 50672.78791614808, "ttft_ms": 33.20241696201265, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c8e95d26a38552eb575a982", "attempt_id": "28070bfdeb46", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:28.762818Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 399.81087495107204, "ttft_ms": 26.320667006075382, "decode_ms": 37.00666700024158}, "total_wall_ms": 50529.06562411226, "ttft_ms": 26.320667006075382, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c43cff88447e21dc42820ca6", "attempt_id": "83fffd4dff2a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:29.252500Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 598, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 368.8224999932572, "ttft_ms": 37.37879206892103, "decode_ms": 60.19529199693352}, "total_wall_ms": 50532.340874080546, "ttft_ms": 37.37879206892103, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a1bb55f3712456ea3249b2ee", "attempt_id": "687f24852679", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:29.742906Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 375.2056249650195, "ttft_ms": 97.71237499080598, "decode_ms": 284.4003749778494}, "total_wall_ms": 50823.55583307799, "ttft_ms": 97.71237499080598, "input_usage": {"lookups": 2}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "935a9e0ab7471ca9463a8be9", "attempt_id": "81dd83ff5bf0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:30.525058Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 388.7675419682637, "ttft_ms": 41.93945892620832, "decode_ms": 25.501957978121936}, "total_wall_ms": 50522.250250098296, "ttft_ms": 41.93945892620832, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fdaa13341ec4c1fc7a664a50", "attempt_id": "4564dff173e6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:31.004874Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 347.77904197108, "ttft_ms": 46.8779579969123, "decode_ms": 60.20466599147767}, "total_wall_ms": 50520.81770810764, "ttft_ms": 46.8779579969123, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4b0a72da9116abb7a58cd292", "attempt_id": "4cbae00fba9c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:31.483534Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 438.1055829580873, "ttft_ms": 132.74558307603002, "decode_ms": 29.905749950557947}, "total_wall_ms": 50666.71066614799, "ttft_ms": 132.74558307603002, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "86ccb38704d3064f3baaa3d8", "attempt_id": "9282b794fa34", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:32.109091Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 375.27050008066, "ttft_ms": 31.685042078606784, "decode_ms": 30.35516699310392}, "total_wall_ms": 50503.16795811523, "ttft_ms": 31.685042078606784, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dcafbac43bd664f1b45a6bf0", "attempt_id": "d71116b0a686", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:32.570125Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 610, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 392.97104196157306, "ttft_ms": 20.669916993938386, "decode_ms": 261.4798340946436}, "total_wall_ms": 50741.15883302875, "ttft_ms": 20.669916993938386, "input_usage": {"lookups": 2}, "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": "755615", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "8e5f423d26733b737903e04c5953aa05727c21f93ccd85f8b611a4116e99673e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d8c8c69a4af7e9f69ba1115c", "attempt_id": "afd6de06a949", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:33.268805Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 535.2573340060189, "ttft_ms": 26.686750003136694, "decode_ms": 289.34999997727573}, "total_wall_ms": 50917.49200015329, "ttft_ms": 26.686750003136694, "input_usage": {"lookups": 2}, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a0e1bed2a7bba1c43b5db418", "attempt_id": "17f49616844e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:34.146408Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 386.0032500233501, "ttft_ms": 21.39091701246798, "decode_ms": 14.135916018858552}, "total_wall_ms": 50487.55504109431, "ttft_ms": 21.39091701246798, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5bb313fe0baaa8c8e9067fd0", "attempt_id": "24957ab88fa2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:34.593469Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 373.5851669916883, "ttft_ms": 37.98000002279878, "decode_ms": 391.83833298739046}, "total_wall_ms": 50869.66866615694, "ttft_ms": 37.98000002279878, "input_usage": {"lookups": 2}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "243879634cd6f5808c9a19ee", "attempt_id": "d9f856eb7ae0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:35.425360Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 600, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 383.60008399467915, "ttft_ms": 38.25420804787427, "decode_ms": 146.9738750020042}, "total_wall_ms": 50634.85579111148, "ttft_ms": 38.25420804787427, "input_usage": {"lookups": 2}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4325045777346143d4d66ce4", "attempt_id": "5f2387c269b6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:36.017749Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 356.34945903439075, "ttft_ms": 36.17658303119242, "decode_ms": 63.83833289146423}, "total_wall_ms": 50522.77837507427, "ttft_ms": 36.17658303119242, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "78bdfb7d93a8049bce587056", "attempt_id": "86decd2add49", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:15:36.498429Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1804, "cache_tensor_bytes": null, "serialized_payload_bytes": 1804, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 50065.194333088584, "hops": {"1": 27020.851625013165, "2": 23044.34270807542}, "receiver_prefill_ms": 523.4457090264186, "ttft_ms": 39.798749960027635, "decode_ms": 24.10999999847263}, "total_wall_ms": 50653.80783309229, "ttft_ms": 39.798749960027635, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "dea86bded5dffd14ea000f98a9df1bfb81a341eb47a7b1b7a9a16db6f1350d87", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0bc3d959927d2f07c82c66cc", "attempt_id": "58292ac12394", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:17.679042Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 501, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 316.7816249188036, "ttft_ms": 36.82729206047952, "decode_ms": 43.52470894809812}, "total_wall_ms": 91024.2117921589, "ttft_ms": 36.82729206047952, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5fe8b93de08af5c2c14cf86d", "attempt_id": "dfc74782bc6d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:18.104332Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 509, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 313.4264999534935, "ttft_ms": 57.36937501933426, "decode_ms": 50.831375061534345}, "total_wall_ms": 91048.67362615187, "ttft_ms": 57.36937501933426, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "be5801a10c8894ec45b01a8a", "attempt_id": "e9af57e10b14", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:18.552632Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 507, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 429.67991600744426, "ttft_ms": 69.11845901049674, "decode_ms": 236.84275010600686}, "total_wall_ms": 91362.62462602463, "ttft_ms": 69.11845901049674, "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": "79.3209", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "00844526dd9401ae600bda6d0889bf28bbf827181a645be64bc8e9566d051361"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e86f2d01d1cc79789fcfa12f", "attempt_id": "def5b4b3ed91", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:19.315768Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 502, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 319.51191602274776, "ttft_ms": 38.238208973780274, "decode_ms": 144.65370902325958}, "total_wall_ms": 91129.41962608602, "ttft_ms": 38.238208973780274, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "147c3369a2f07b9d622bcb49", "attempt_id": "5c8447ddfb8b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:19.843084Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 509, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 369.77054097224027, "ttft_ms": 42.94791701249778, "decode_ms": 107.75450000073761}, "total_wall_ms": 91147.5452920422, "ttft_ms": 42.94791701249778, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f9163fe2e772666076733c2c", "attempt_id": "de9a9c7499be", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:20.389294Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 511, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 402.2642499767244, "ttft_ms": 32.329041976481676, "decode_ms": 38.245624979026616}, "total_wall_ms": 91099.87537609413, "ttft_ms": 32.329041976481676, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "00e890b12422c5602232a532", "attempt_id": "a0ef9e86a1ce", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:20.888125Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 502, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 342.5106250215322, "ttft_ms": 36.23983298894018, "decode_ms": 37.572499946691096}, "total_wall_ms": 91043.19029208273, "ttft_ms": 36.23983298894018, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f0af138d1553529db60b4f67", "attempt_id": "4c9f5798d962", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:21.327575Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 504, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 315.29450009111315, "ttft_ms": 62.787957955151796, "decode_ms": 63.16308304667473}, "total_wall_ms": 91068.32625111565, "ttft_ms": 62.787957955151796, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a3350788fec28619273ebdb1", "attempt_id": "64fb6b61b3f6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:21.792236Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 519, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 498.69083403609693, "ttft_ms": 36.47754201665521, "decode_ms": 23.786292062141}, "total_wall_ms": 91186.0227510333, "ttft_ms": 36.47754201665521, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cdcb0d4a32d48ae431bdd907", "attempt_id": "76638bf421bc", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:22.379587Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 513, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 328.4309159498662, "ttft_ms": 31.497249961830676, "decode_ms": 259.7171670058742}, "total_wall_ms": 91246.73850112595, "ttft_ms": 31.497249961830676, "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": "10190005", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "c37bcd57e6e0d2e6cdf5cd63d515941554ca0506c135932416bc1f3bb679fa86"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1a17d5566b765499d2429788", "attempt_id": "aca7de4a2c4c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:23.024008Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 520, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 352.2915419889614, "ttft_ms": 58.1300409976393, "decode_ms": 107.29358298704028}, "total_wall_ms": 91145.00379317906, "ttft_ms": 58.1300409976393, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "44e34d848aa03463d876e1b2", "attempt_id": "4aca1b5186e7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:23.569793Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 514, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 441.43000000622123, "ttft_ms": 23.474625078961253, "decode_ms": 215.73850000277162}, "total_wall_ms": 91308.00708406605, "ttft_ms": 23.474625078961253, "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": "744337", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "871006e2b4296b009ebd6d7e4fab0d22ce7b550aa2a6c3539f54474eadc9864b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "24d858d7aa74b47afacb7745", "attempt_id": "549b198d8d2e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:24.277542Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 509, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 320.695957983844, "ttft_ms": 22.22233300562948, "decode_ms": 24.289917084388435}, "total_wall_ms": 90994.38541708514, "ttft_ms": 22.22233300562948, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "108aac7d375edb2aff97b852", "attempt_id": "6d699f610c2f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:24.670267Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 513, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 337.67849998548627, "ttft_ms": 62.45874997694045, "decode_ms": 93.84895802941173}, "total_wall_ms": 91121.13279209007, "ttft_ms": 62.45874997694045, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fbbc62a7d4a1556a0224c135", "attempt_id": "b15d9a0bde43", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:25.190961Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 516, "message_tokens": 421, "message_utf8_bytes": 1484, "cache_tensor_bytes": null, "serialized_payload_bytes": 1484, "side_information_bytes": 0, "fetched_bytes": 3324, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 90626.24550110195, "hops": {"1": 27020.851625013165, "2": 23044.34270807542, "3": 19254.31083398871, "4": 21306.740334024653}, "receiver_prefill_ms": 431.9628750672564, "ttft_ms": 24.203874985687435, "decode_ms": 43.63395902328193}, "total_wall_ms": 91126.70416815672, "ttft_ms": 24.203874985687435, "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": {"ledger": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 19, "retention_rate": 0.296875, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1813}, "2": {"facts_total": 64, "facts_retained": 20, "retention_rate": 0.3125, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.25, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1804}, "3": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}, "4": {"facts_total": 64, "facts_retained": 22, "retention_rate": 0.34375, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.35714285714285715, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1484}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "6a280c6c2331b55c7c109ec9dc50404550db655f2d11b13bf86efa9b3e1cb463", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "97925e87ae0386b18489e0c3", "attempt_id": "50237e903173", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:52.464954Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 372.41058296058327, "ttft_ms": 133.88095807749778, "decode_ms": 36.5722089773044}, "total_wall_ms": 27289.679958019406, "ttft_ms": 133.88095807749778, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d834fbc27245cbd458f38372", "attempt_id": "c2208b9112ea", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:53.031590Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 279.3596249539405, "ttft_ms": 21.748999948613346, "decode_ms": 35.02345806919038}, "total_wall_ms": 27083.238000050187, "ttft_ms": 21.748999948613346, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "de9e679ad7acec566f029b7c", "attempt_id": "b61d1b0227ba", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:53.392722Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 441, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 284.6724590053782, "ttft_ms": 31.080750050023198, "decode_ms": 42.81412495765835}, "total_wall_ms": 27105.419207946397, "ttft_ms": 31.080750050023198, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4bca5742eaaeb2bcf117f2aa", "attempt_id": "5d5f6aec9691", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:53.777017Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 250.13666693121195, "ttft_ms": 52.682625013403594, "decode_ms": 301.05104204267263}, "total_wall_ms": 27350.92454100959, "ttft_ms": 52.682625013403594, "input_usage": {"lookups": 0}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "77703830b4ef3128fec35f8b", "attempt_id": "799f4736b3c3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:54.405219Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 359.8788329400122, "ttft_ms": 40.97254201769829, "decode_ms": 26.214416953735054}, "total_wall_ms": 27173.946416005492, "ttft_ms": 40.97254201769829, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "df3405ae70c4de6036aa0853", "attempt_id": "d52b18aec881", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:54.855317Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 445, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 280.7114589959383, "ttft_ms": 40.998667012900114, "decode_ms": 29.205165919847786}, "total_wall_ms": 27098.002457991242, "ttft_ms": 40.998667012900114, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "135c8e20ab329d7586e04dd9", "attempt_id": "1d691a65da49", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:55.229847Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 436, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 258.1689580110833, "ttft_ms": 56.5575419459492, "decode_ms": 59.29125007241964}, "total_wall_ms": 27120.92604103964, "ttft_ms": 56.5575419459492, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c3a028206aa766c46499984", "attempt_id": "af51a20ccb18", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:55.629972Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 438, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 284.05670798383653, "ttft_ms": 94.08091695513576, "decode_ms": 135.12916606850922}, "total_wall_ms": 27260.410791030154, "ttft_ms": 94.08091695513576, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f42cbaea23fc6edeb00ac8e0", "attempt_id": "58dce6b5698c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:56.166643Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 453, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 295.0127919903025, "ttft_ms": 35.581458010710776, "decode_ms": 23.571291007101536}, "total_wall_ms": 27101.123749045655, "ttft_ms": 35.581458010710776, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfd4a112e3348bc21f9d8536", "attempt_id": "41850a169fdc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:56.545027Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 284.07120797783136, "ttft_ms": 38.345832959748805, "decode_ms": 24.728541960939765}, "total_wall_ms": 27094.10524996929, "ttft_ms": 38.345832959748805, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d074f7dfb7730b04327b278c", "attempt_id": "2e9fa6a9439f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:56.917174Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 454, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 271.51675010100007, "ttft_ms": 53.51787502877414, "decode_ms": 37.86995797418058}, "total_wall_ms": 27109.752124059014, "ttft_ms": 53.51787502877414, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "add3f195a893af9a3d54aaca", "attempt_id": "340b0e668992", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:57.303826Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 448, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 277.68041694071144, "ttft_ms": 94.22037506010383, "decode_ms": 328.61399999819696}, "total_wall_ms": 27447.520291083492, "ttft_ms": 94.22037506010383, "input_usage": {"lookups": 0}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2e967f6f71623d99fd061ad3", "attempt_id": "b76ec690881e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:58.028975Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 443, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 288.9045000774786, "ttft_ms": 22.86699996329844, "decode_ms": 176.40054097864777}, "total_wall_ms": 27235.02883198671, "ttft_ms": 22.86699996329844, "input_usage": {"lookups": 0}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "603694ece680e14da83b8b1e", "attempt_id": "6db84adaad42", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:58.544073Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 447, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 296.1954170605168, "ttft_ms": 48.4462499152869, "decode_ms": 55.72541698347777}, "total_wall_ms": 27147.335456917062, "ttft_ms": 48.4462499152869, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5e65242ac83b0beaffc8f990", "attempt_id": "effe0faf2298", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:16:58.968926Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 450, "message_tokens": 354, "message_utf8_bytes": 832, "cache_tensor_bytes": null, "serialized_payload_bytes": 832, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26746.30691600032, "hops": {"1": 26746.30691600032}, "receiver_prefill_ms": 457.12979102972895, "ttft_ms": 35.45933298300952, "decode_ms": 43.70583291165531}, "total_wall_ms": 27283.183791092597, "ttft_ms": 35.45933298300952, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "12fceaa4bed42c6e4a8870fc960d8a8de8b964a036bae09a5048ced95310d3f0", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "92a4e2c99a26e7d4cf7e3a7f", "attempt_id": "d7b268ef3020", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:15.147536Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 416, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 286.3557090749964, "ttft_ms": 36.59249993506819, "decode_ms": 23.184292018413544}, "total_wall_ms": 42708.5580410203, "ttft_ms": 36.59249993506819, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5471d24b8aaf94f4adf76fd0", "attempt_id": "cb4999fec1f1", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:15.518692Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 285.93466710299253, "ttft_ms": 37.08270797505975, "decode_ms": 26.987249962985516}, "total_wall_ms": 42712.372000096366, "ttft_ms": 37.08270797505975, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8e026cc350c06ed456db926d", "attempt_id": "cd61478a8a30", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:15.892067Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 422, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 254.16591705288738, "ttft_ms": 56.24304199591279, "decode_ms": 43.177250074222684}, "total_wall_ms": 42716.05275000911, "ttft_ms": 56.24304199591279, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2d6ef2f2ee398e2539d72aa9", "attempt_id": "0b09e4099e97", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:16.269453Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 417, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 371.9577919691801, "ttft_ms": 78.96708301268518, "decode_ms": 175.87499995715916}, "total_wall_ms": 42989.37533306889, "ttft_ms": 78.96708301268518, "input_usage": {"lookups": 2}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2bc0540e4d58def8170e9cda", "attempt_id": "ff0faf010289", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:16.924190Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 288.17062510643154, "ttft_ms": 25.84274997934699, "decode_ms": 221.88516706228256}, "total_wall_ms": 42898.55037408415, "ttft_ms": 25.84274997934699, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "898,925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "e9993a05a91873daaff830fc3ab8afceb2667baf488f2814ad8d537c1dba9101"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "772758790d83aaf15836f78e", "attempt_id": "3b86df3bb02f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:17.485879Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 426, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 320.02124993596226, "ttft_ms": 38.65233401302248, "decode_ms": 58.6811660323292}, "total_wall_ms": 42780.08929104544, "ttft_ms": 38.65233401302248, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c50ba29b3c7628a0e99c0b2", "attempt_id": "5347467b164e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:17.926790Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 417, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 421.12991597969085, "ttft_ms": 37.043208023533225, "decode_ms": 26.4788749627769}, "total_wall_ms": 42847.06658299547, "ttft_ms": 37.043208023533225, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2c0b6775ac1c5ab4f9aa37a4", "attempt_id": "823726bfe8fb", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:18.435448Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 419, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 274.65491695329547, "ttft_ms": 29.90616695024073, "decode_ms": 39.19616702478379}, "total_wall_ms": 42706.241416046396, "ttft_ms": 29.90616695024073, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8efa9c7119c9df6d81cf3707", "attempt_id": "0a42fb765afd", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:18.805270Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 434, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 283.3426669239998, "ttft_ms": 20.633084001019597, "decode_ms": 19.365041982382536}, "total_wall_ms": 42685.66608300898, "ttft_ms": 20.633084001019597, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "552e16b1826c324fe1b37da3", "attempt_id": "dbb287f03758", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:19.154894Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 428, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 302.26679099723697, "ttft_ms": 61.12066702917218, "decode_ms": 60.02337497193366}, "total_wall_ms": 42785.77854100149, "ttft_ms": 61.12066702917218, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "910aa3a7ea77aebfb6020b65", "attempt_id": "0aa1b60e44f2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:19.602797Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 399.5379590196535, "ttft_ms": 34.76537496317178, "decode_ms": 43.60795801039785}, "total_wall_ms": 42840.58458299842, "ttft_ms": 34.76537496317178, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f8ae66f49e5ae86123765ae3", "attempt_id": "2e9a9b3653f9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:20.103848Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 429, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 292.31241706293076, "ttft_ms": 37.759916041977704, "decode_ms": 203.6084580468014}, "total_wall_ms": 42896.12737507559, "ttft_ms": 37.759916041977704, "input_usage": {"lookups": 2}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6594dc89cda6d3dd3a282461", "attempt_id": "b4bd88a8f8fc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:20.665160Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 303.5563339944929, "ttft_ms": 54.39266702160239, "decode_ms": 358.06812497321516}, "total_wall_ms": 43078.63441610243, "ttft_ms": 54.39266702160239, "input_usage": {"lookups": 2}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9fd64ed2fb0d732d99ad940b", "attempt_id": "aa878be0a2ce", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:21.407795Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 428, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 276.3891250360757, "ttft_ms": 39.26470805890858, "decode_ms": 25.104832951910794}, "total_wall_ms": 42703.26604205184, "ttft_ms": 39.26470805890858, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2f9c189fc3165ce05cd9cde2", "attempt_id": "4ea3e5ff2464", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:21.773010Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 431, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 1108, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 42361.81558307726, "hops": {"1": 26746.30691600032, "2": 15615.508667076938}, "receiver_prefill_ms": 277.3655420169234, "ttft_ms": 30.533958110027015, "decode_ms": 33.034457941539586}, "total_wall_ms": 42703.48379202187, "ttft_ms": 30.533958110027015, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "86ae90b140e95a1ae29161b5", "attempt_id": "59a1f97bde50", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:52.953373Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 416, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 303.09108295477927, "ttft_ms": 38.91516593284905, "decode_ms": 24.712457903660834}, "total_wall_ms": 73538.99512405042, "ttft_ms": 38.91516593284905, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1e262221803aaea89556ca4e", "attempt_id": "c3cd0fb63a6c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:53.343613Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 288.1236670073122, "ttft_ms": 36.55299998354167, "decode_ms": 34.0587500249967}, "total_wall_ms": 73530.9816240333, "ttft_ms": 36.55299998354167, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2f846a4f02e71296ade6f31d", "attempt_id": "75a00ad7081b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:53.725657Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 422, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 298.9596250699833, "ttft_ms": 40.38320807740092, "decode_ms": 56.20520806405693}, "total_wall_ms": 73567.76512402575, "ttft_ms": 40.38320807740092, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a22291a09f0b47e807f84a33", "attempt_id": "69d98038ac12", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:54.146381Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 417, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 419.6435830090195, "ttft_ms": 38.89741702005267, "decode_ms": 168.43320801854134}, "total_wall_ms": 73799.46779098827, "ttft_ms": 38.89741702005267, "input_usage": {"lookups": 4}, "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": "Haruki Achterberg", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "f8f4436d5607232dde4f6986ce4cf4ad38bcb02f807312d9c5078948d1ac80b3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "87e7cd351a52c0a7c19f000e", "attempt_id": "a227767a1cdb", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:54.796319Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 279.5157090295106, "ttft_ms": 43.241666979156435, "decode_ms": 255.2338750101626}, "total_wall_ms": 73750.50487404224, "ttft_ms": 43.241666979156435, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "898,925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "e9993a05a91873daaff830fc3ab8afceb2667baf488f2814ad8d537c1dba9101"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3e774f9ceb5b5edd3f7530b1", "attempt_id": "4b2e52e42c97", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:55.398625Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 426, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 275.17337491735816, "ttft_ms": 113.06208395399153, "decode_ms": 143.54583295062184}, "total_wall_ms": 73704.16000008117, "ttft_ms": 113.06208395399153, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "834692737b5ca38c05856538", "attempt_id": "916d7d20281f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:55.955057Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 417, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 271.5991250006482, "ttft_ms": 40.382082923315465, "decode_ms": 28.887958964332938}, "total_wall_ms": 73513.26495804824, "ttft_ms": 40.382082923315465, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd4aa7709fa0ce62217a5205", "attempt_id": "13e3d6241fd0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:56.320218Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 419, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 279.24245805479586, "ttft_ms": 40.42354098055512, "decode_ms": 31.26445901580155}, "total_wall_ms": 73523.10008194763, "ttft_ms": 40.42354098055512, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9badc895f2db0618d7c5aaff", "attempt_id": "036484d692dd", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:56.694788Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 434, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 246.6167500242591, "ttft_ms": 71.73316599801183, "decode_ms": 68.62712500151247}, "total_wall_ms": 73559.33041602839, "ttft_ms": 71.73316599801183, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6d8a27c0f618c11ae4c7aeb0", "attempt_id": "15373be7dcd5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:57.104670Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 428, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 319.6262080455199, "ttft_ms": 133.9871670352295, "decode_ms": 42.72737505380064}, "total_wall_ms": 73668.54087403044, "ttft_ms": 133.9871670352295, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bf4ff161e132b31be3da29b5", "attempt_id": "77e9e24fe0a5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:57.624631Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 435, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 281.536583090201, "ttft_ms": 43.633875087834895, "decode_ms": 24.00591701734811}, "total_wall_ms": 73521.54691610485, "ttft_ms": 43.633875087834895, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "37feb17ce3b7b3aa0c33719e", "attempt_id": "e7c5e5f7c47f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:57.997688Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 429, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 278.7731249118224, "ttft_ms": 20.65262501128018, "decode_ms": 181.7169999703765}, "total_wall_ms": 73653.66383211222, "ttft_ms": 20.65262501128018, "input_usage": {"lookups": 4}, "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": "898925", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "6be0abe995dc52d82015157d72ecc12116d3bfb521379f2d9818e289666dbee4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a105d5f71dff3d069d7ba49e", "attempt_id": "87401506ebc5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:58.504753Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 424, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 278.5967090167105, "ttft_ms": 35.142250009812415, "decode_ms": 371.6344579588622}, "total_wall_ms": 73857.88599902298, "ttft_ms": 35.142250009812415, "input_usage": {"lookups": 4}, "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": "21.14", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "823f6178dbf8cdf8c128dbc87330293f2b298358418da7168fb7f865128fa176"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7a0b1856a28a1471695b90de", "attempt_id": "121ddab665b3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:59.214866Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 428, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 283.9384169783443, "ttft_ms": 32.69920905586332, "decode_ms": 38.766250014305115}, "total_wall_ms": 73528.04524998646, "ttft_ms": 32.69920905586332, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dc326fa361b51f4486fe3ce7", "attempt_id": "2b6f5ec0e681", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:17:59.597256Z", "task_family": "synthetic", "document_id": "syn_02001_8000", "document_hash": "acefbded40b5210f20d4288aa9a950b81e67729e53809c2fd0da03d1517011ae", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02001_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5724, "query_tokens": 431, "message_tokens": 336, "message_utf8_bytes": 851, "cache_tensor_bytes": null, "serialized_payload_bytes": 851, "side_information_bytes": 0, "fetched_bytes": 2146, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73171.69566603843, "hops": {"1": 26746.30691600032, "2": 15615.508667076938, "3": 15459.52087501064, "4": 15350.359207950532}, "receiver_prefill_ms": 283.9437499642372, "ttft_ms": 21.464833058416843, "decode_ms": 35.8174170833081}, "total_wall_ms": 73513.5529161198, "ttft_ms": 21.464833058416843, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "ledger_by_hop": {"1": {"facts_total": 64, "facts_retained": 16, "retention_rate": 0.25, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.14285714285714285, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 832}, "2": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "3": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}, "4": {"facts_total": 64, "facts_retained": 15, "retention_rate": 0.234375, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.07142857142857142, "status": 0.75, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 851}}}, "reference_key_id": "keys/synthetic/syn_02001_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/E4_pilot.log", "payload_hash": "fdfd6d3ceca4aab9e3f57a338c4fe269e5ba85a19b238c9dc8b2bbdd1c6d19fc", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f7b4717b6fd6bab8cbca3d03", "attempt_id": "8111e2b5de1f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:23.785221Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 358.2646249560639, "ttft_ms": 69.14720800705254, "decode_ms": 59.78491704445332}, "total_wall_ms": 24213.110166951083, "ttft_ms": 69.14720800705254, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c688848dc14391472e8f4cd6", "attempt_id": "2c13c40fa000", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:24.295213Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 543.5404169838876, "ttft_ms": 21.858458989299834, "decode_ms": 44.75345800165087}, "total_wall_ms": 24335.921832942404, "ttft_ms": 21.858458989299834, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "76c5abd3ea5a3e613709f401", "attempt_id": "9f9ca8fad901", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:24.926473Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 391.35154196992517, "ttft_ms": 20.914750057272613, "decode_ms": 269.8150419164449}, "total_wall_ms": 24408.006374025717, "ttft_ms": 20.914750057272613, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35.5234", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "005768a0245176a932a31dd800f8ef72b75c295d00e4f081a98f2978ca880367"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d0a8e8ca11024aab4d5b3d67", "attempt_id": "2c0058af683e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:25.630390Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 427.55200003739446, "ttft_ms": 139.40033304970711, "decode_ms": 253.35720798466355}, "total_wall_ms": 24546.082833083346, "ttft_ms": 139.40033304970711, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eed037ec06c0780a0e95d13c", "attempt_id": "9fa375925525", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:26.471423Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 404.0440840180963, "ttft_ms": 34.58037495147437, "decode_ms": 349.37937499489635}, "total_wall_ms": 24513.855583034456, "ttft_ms": 34.58037495147437, "input_usage": {"lookups": 0}, "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": "1,844,550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "c3a824f75c26a69ab7e2a0dfde5b194c586a879a105eff635ab136cdb1a4fd6c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "60589ff33e8f3d2468a84865", "attempt_id": "7d199e0bb854", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:27.280838Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 509.46537498384714, "ttft_ms": 26.28616604488343, "decode_ms": 37.95404196716845}, "total_wall_ms": 24299.532582983375, "ttft_ms": 26.28616604488343, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5a5b7a526784d38fca2135f6", "attempt_id": "b931c3c36e59", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:27.875447Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 392.9324160562828, "ttft_ms": 32.91129192803055, "decode_ms": 30.079666059464216}, "total_wall_ms": 24181.576375034638, "ttft_ms": 32.91129192803055, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "240aee8dfea8146d4f80eb47", "attempt_id": "49d5bb6eb4d3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:28.352402Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 374.29874995723367, "ttft_ms": 21.43383293878287, "decode_ms": 58.886790997348726}, "total_wall_ms": 24180.795416003093, "ttft_ms": 21.43383293878287, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b7f492087da097bc1c698b54", "attempt_id": "23788c64d52f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:28.827309Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 609, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 387.8371670143679, "ttft_ms": 91.61391702946275, "decode_ms": 142.54004100803286}, "total_wall_ms": 24347.969957976602, "ttft_ms": 91.61391702946275, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2577cb7316938f27e2fe9254", "attempt_id": "dc9432527479", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:29.471242Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 379.97454195283353, "ttft_ms": 37.167917005717754, "decode_ms": 236.13850004039705}, "total_wall_ms": 24379.17429197114, "ttft_ms": 37.167917005717754, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "692ce73e1cb6a268fd1d2a27", "attempt_id": "166f77f3c7a1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:30.146830Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 374.2560000391677, "ttft_ms": 52.59233294054866, "decode_ms": 67.5543750403449}, "total_wall_ms": 24220.21283302456, "ttft_ms": 52.59233294054866, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a8101c4cefe5cfcd9deb1d95", "attempt_id": "c385099cb313", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:30.662329Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 567.5797500880435, "ttft_ms": 35.983540932647884, "decode_ms": 251.63070904091}, "total_wall_ms": 24580.93854098115, "ttft_ms": 35.983540932647884, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5464.31", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "754be52ae0e9ea0a9d6f318d0a6289c2971ea97dc372d8c4a25cd50e97b97540"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "277029145369f087fcd9b5f5", "attempt_id": "2418a54e0034", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:31.539651Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 384.9255000241101, "ttft_ms": 21.955000003799796, "decode_ms": 22.07920898217708}, "total_wall_ms": 24154.704582993872, "ttft_ms": 21.955000003799796, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "daf6c86f1d5e35f121bd7111", "attempt_id": "10546ea1a5ca", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:31.988838Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 361.57154105603695, "ttft_ms": 48.85470902081579, "decode_ms": 43.22983301244676}, "total_wall_ms": 24179.36933308374, "ttft_ms": 48.85470902081579, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4d5fe51005e7f9b3462f4e2e", "attempt_id": "bdc286c4aa71", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:32.464453Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23725.01320799347, "hops": {"1": 23725.01320799347}, "receiver_prefill_ms": 519.1463329829276, "ttft_ms": 34.42416596226394, "decode_ms": 34.117499948479235}, "total_wall_ms": 24313.540207920596, "ttft_ms": 34.42416596226394, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7fbf000ccd80cfad6b82a076", "attempt_id": "106c5134dfcb", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:55.875865Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 569.7251659585163, "ttft_ms": 22.42237504106015, "decode_ms": 38.75383303966373}, "total_wall_ms": 47158.57787407003, "ttft_ms": 22.42237504106015, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62dee344d784fb93fb006ff2", "attempt_id": "ad3e702168f6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:56.527651Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 405.73641599621624, "ttft_ms": 39.28579099010676, "decode_ms": 31.27112495712936}, "total_wall_ms": 47003.80741502158, "ttft_ms": 39.28579099010676, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d24059c987c59bd08de5b74e", "attempt_id": "877a182188e3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:57.023453Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 352.5982089340687, "ttft_ms": 53.09258308261633, "decode_ms": 504.2751660803333}, "total_wall_ms": 47437.57466610987, "ttft_ms": 53.09258308261633, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35.5234", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "005768a0245176a932a31dd800f8ef72b75c295d00e4f081a98f2978ca880367"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f08706eba9c6968ca4a87c5c", "attempt_id": "aeecac46fae7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:57.952732Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 399.4135420070961, "ttft_ms": 34.28066603373736, "decode_ms": 257.34200002625585}, "total_wall_ms": 47218.63329003099, "ttft_ms": 34.28066603373736, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "be62ddd269877f9db8beee38", "attempt_id": "87f2b49fc781", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:58.662549Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 376.50837493129075, "ttft_ms": 52.05299996305257, "decode_ms": 507.6943750027567}, "total_wall_ms": 47464.12033215165, "ttft_ms": 52.05299996305257, "input_usage": {"lookups": 0}, "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": "1,844,550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "c3a824f75c26a69ab7e2a0dfde5b194c586a879a105eff635ab136cdb1a4fd6c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a5e89327e6e2ab76a903481e", "attempt_id": "a02016430463", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:18:59.619763Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 400.6038340739906, "ttft_ms": 33.59070897568017, "decode_ms": 31.192292110063136}, "total_wall_ms": 46993.04095702246, "ttft_ms": 33.59070897568017, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9ab03c86145ad0b0b9c57871", "attempt_id": "d63d8c1f6954", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:00.106961Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 377.54387501627207, "ttft_ms": 58.98770794738084, "decode_ms": 48.40412503108382}, "total_wall_ms": 47012.561958050355, "ttft_ms": 58.98770794738084, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5c5de7fb30d7e218380a4797", "attempt_id": "7055f04b82b6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:00.611145Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 552.6042920537293, "ttft_ms": 42.20362496562302, "decode_ms": 26.90687496215105}, "total_wall_ms": 47149.370832019486, "ttft_ms": 42.20362496562302, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f18f356ac1deac00e3e38c90", "attempt_id": "32ec544e98a1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:01.254653Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 609, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 377.25145800504833, "ttft_ms": 23.53562496136874, "decode_ms": 35.346666001714766}, "total_wall_ms": 46963.76354107633, "ttft_ms": 23.53562496136874, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ff2bc9898d9e05f05cc4e027", "attempt_id": "de10e585bde3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:01.710523Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 379.0871249511838, "ttft_ms": 20.126374904066324, "decode_ms": 428.99924993980676}, "total_wall_ms": 47355.8507070411, "ttft_ms": 20.126374904066324, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e949f7523f3cfac962753a2d", "attempt_id": "d54ab797d377", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:02.559738Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 469.5638329721987, "ttft_ms": 36.46191698499024, "decode_ms": 29.948625015094876}, "total_wall_ms": 47063.55870713014, "ttft_ms": 36.46191698499024, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a6f689ff8d7b842911713a44", "attempt_id": "4ad977a0a73f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:03.117797Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 388.5044170310721, "ttft_ms": 39.19579100329429, "decode_ms": 270.72545897681266}, "total_wall_ms": 47225.989249069244, "ttft_ms": 39.19579100329429, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5464.31", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "754be52ae0e9ea0a9d6f318d0a6289c2971ea97dc372d8c4a25cd50e97b97540"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fbbf0f5c8d4d9d232402d7bd", "attempt_id": "94aade151570", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:03.836859Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 571.0587499197572, "ttft_ms": 35.70308303460479, "decode_ms": 30.61391704250127}, "total_wall_ms": 47165.184041135944, "ttft_ms": 35.70308303460479, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "77873f4bedb4b278250fa687", "attempt_id": "217f0ef8e586", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:04.495883Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 382.9038329422474, "ttft_ms": 21.957915974780917, "decode_ms": 42.113584000617266}, "total_wall_ms": 46974.62683310732, "ttft_ms": 21.957915974780917, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "74f5a425ea2cd322084e61ca", "attempt_id": "6de12bcbdb22", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:04.963512Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46526.87412407249, "hops": {"1": 23725.01320799347, "2": 22801.860916079022}, "receiver_prefill_ms": 375.99191605113447, "ttft_ms": 38.58724993187934, "decode_ms": 39.92995910812169}, "total_wall_ms": 46982.07991605159, "ttft_ms": 38.58724993187934, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "15aa295b2751148c84897123", "attempt_id": "e5bc87c1271e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:50.985983Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 400.35783301573247, "ttft_ms": 92.33491600025445, "decode_ms": 128.466916969046}, "total_wall_ms": 92693.2286650408, "ttft_ms": 92.33491600025445, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ae252823a1388718a2029675", "attempt_id": "81dc10f7b44e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:51.628295Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 384.35750000644475, "ttft_ms": 38.863791967742145, "decode_ms": 39.97562499716878}, "total_wall_ms": 92535.16483202111, "ttft_ms": 38.863791967742145, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8b9e2f74532e41373a7721ec", "attempt_id": "d3bdd9d679a3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:52.111803Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 384.7306250827387, "ttft_ms": 40.80141708254814, "decode_ms": 308.7876250501722}, "total_wall_ms": 92806.35020707268, "ttft_ms": 40.80141708254814, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35.5234", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "005768a0245176a932a31dd800f8ef72b75c295d00e4f081a98f2978ca880367"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2a4dbd30d081cbaa63227b4a", "attempt_id": "884f132c651f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:52.866775Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 529.4262500246987, "ttft_ms": 36.29933309275657, "decode_ms": 249.05987503007054}, "total_wall_ms": 92886.76387409214, "ttft_ms": 36.29933309275657, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "80b7166798e8e496001e9811", "attempt_id": "ffecc4ea62d4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:53.701620Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 386.08862494584173, "ttft_ms": 27.980832965113223, "decode_ms": 525.2292080549523}, "total_wall_ms": 93011.20983203873, "ttft_ms": 27.980832965113223, "input_usage": {"lookups": 0}, "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": "1,844,550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "c3a824f75c26a69ab7e2a0dfde5b194c586a879a105eff635ab136cdb1a4fd6c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dbfceb73bc892d992dcf2a5c", "attempt_id": "8bcd60049334", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:54.661653Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 421.5734170284122, "ttft_ms": 35.31199996359646, "decode_ms": 43.050749925896525}, "total_wall_ms": 92571.8540820526, "ttft_ms": 35.31199996359646, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "535ae8e9ead8ef629af2cc1d", "attempt_id": "f6755fcfd05c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:55.182366Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 405.4134580073878, "ttft_ms": 34.42549996543676, "decode_ms": 31.237541930750012}, "total_wall_ms": 92542.94429102447, "ttft_ms": 34.42549996543676, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c8f5007db1df8c8160854054", "attempt_id": "f6a39b3d1192", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:55.674056Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 362.1021250728518, "ttft_ms": 53.44712489750236, "decode_ms": 52.28716600686312}, "total_wall_ms": 92540.0522480486, "ttft_ms": 53.44712489750236, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "141a1d1e18e4fd633bfc3443", "attempt_id": "92d736cac0ed", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:56.162576Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 609, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 549.1457500029355, "ttft_ms": 29.9294589785859, "decode_ms": 39.93395809084177}, "total_wall_ms": 92691.04658206925, "ttft_ms": 29.9294589785859, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b9e27dfd94270ced6e19612f", "attempt_id": "de6e581e78fd", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:56.801678Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 400.8960829814896, "ttft_ms": 34.33979209512472, "decode_ms": 241.81195790879428}, "total_wall_ms": 92749.17520699091, "ttft_ms": 34.33979209512472, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2fc67c9be2de799abe522c75", "attempt_id": "02de93bf9d5d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:57.500222Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 471.8437909614295, "ttft_ms": 133.27154202852398, "decode_ms": 20.833084010519087}, "total_wall_ms": 92698.0137899518, "ttft_ms": 133.27154202852398, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62a0dd301354daf0afbf657a", "attempt_id": "5131d1e2329e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:58.146261Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 389.1037499997765, "ttft_ms": 37.972207996062934, "decode_ms": 242.2144579468295}, "total_wall_ms": 92741.30245705601, "ttft_ms": 37.972207996062934, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5464.31", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "754be52ae0e9ea0a9d6f318d0a6289c2971ea97dc372d8c4a25cd50e97b97540"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8a87e58ef5998e32ff37a5f3", "attempt_id": "f2d7b94c5fd2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:58.836029Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 352.0441669970751, "ttft_ms": 35.74920794926584, "decode_ms": 56.54308293014765}, "total_wall_ms": 92516.22812403366, "ttft_ms": 35.74920794926584, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c5e631e714eb768e34e15dd3", "attempt_id": "ea0c2a26e03c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:59.301439Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 554.8922079615295, "ttft_ms": 25.278209010139108, "decode_ms": 36.09087492804974}, "total_wall_ms": 92688.21616610512, "ttft_ms": 25.278209010139108, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d70d0706144fea8e206a63ca", "attempt_id": "d5e03e2e1b17", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:19:59.937400Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1171, "cache_tensor_bytes": null, "serialized_payload_bytes": 1171, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92071.19358202908, "hops": {"1": 23725.01320799347, "2": 22801.860916079022, "3": 22925.79150001984, "4": 22618.52795793675}, "receiver_prefill_ms": 397.19087502453476, "ttft_ms": 33.247165963985026, "decode_ms": 31.379208085127175}, "total_wall_ms": 92533.71287300251, "ttft_ms": 33.247165963985026, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "3": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}, "4": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1171}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "d089f2bb50b7fd1147ae7cc9bf57964c0558bb8bc1a7196b2f63d0c24f6c5d9f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5d07130d0c4887905c4abc5d", "attempt_id": "82e7ea3c431c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:24.143321Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 435, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 289.8689160356298, "ttft_ms": 36.47470904979855, "decode_ms": 24.225250002928078}, "total_wall_ms": 24072.636917000636, "ttft_ms": 36.47470904979855, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0be83922a055aee6aebd2238", "attempt_id": "62032c0ddb75", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:24.513897Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 441, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 294.5528329582885, "ttft_ms": 60.27554196771234, "decode_ms": 51.691000000573695}, "total_wall_ms": 24128.703249967657, "ttft_ms": 60.27554196771234, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "67099c6c938ff0b0f5307587", "attempt_id": "8105adc5ba9f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:24.941344Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 439, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 422.96687501948327, "ttft_ms": 32.77333301957697, "decode_ms": 39.569791057147086}, "total_wall_ms": 24217.3837920418, "ttft_ms": 32.77333301957697, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c3036f04a502f308a309ca70", "attempt_id": "5db6b91e119c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:25.459578Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 280.08383291307837, "ttft_ms": 40.40795797482133, "decode_ms": 24.29724996909499}, "total_wall_ms": 24066.738957888447, "ttft_ms": 40.40795797482133, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "33ce9fefeacb3774812622ed", "attempt_id": "d909aa771e92", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:25.824948Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 291.8473749887198, "ttft_ms": 23.7047920236364, "decode_ms": 532.4121250305325}, "total_wall_ms": 24569.961500004865, "ttft_ms": 23.7047920236364, "input_usage": {"lookups": 0}, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "026f49e1404415ff73812ded", "attempt_id": "c4367572b760", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:26.692243Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 338.9572500018403, "ttft_ms": 33.99799996986985, "decode_ms": 341.10670804511756}, "total_wall_ms": 24435.99795899354, "ttft_ms": 33.99799996986985, "input_usage": {"lookups": 0}, "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-79U-1063", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "0de04aed6b20f3db39c86fb65475b79b59798e252a9c06c61105f9cdf5b96aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "046388b2e5ca1bb0ac195c0e", "attempt_id": "750b68f6775e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:27.425881Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 267.8971668938175, "ttft_ms": 54.83995901886374, "decode_ms": 59.414333081804216}, "total_wall_ms": 24104.228500043973, "ttft_ms": 54.83995901886374, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cd60ebced49e6acf93d2750b", "attempt_id": "37ad23716166", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:27.827359Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 436, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 420.5334580037743, "ttft_ms": 52.76374996174127, "decode_ms": 28.121375013142824}, "total_wall_ms": 24223.347624996677, "ttft_ms": 52.76374996174127, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a2e88a63103ef274b766ccd6", "attempt_id": "e5355d023ffd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:28.350898Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 451, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 302.8691249201074, "ttft_ms": 26.336750015616417, "decode_ms": 36.67170798871666}, "total_wall_ms": 24087.843999965116, "ttft_ms": 26.336750015616417, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "457c6ac338f8f5ef69e2d0c2", "attempt_id": "4cb8639074ce", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:28.736076Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 289.9117920314893, "ttft_ms": 18.778624944388866, "decode_ms": 403.1904579605907}, "total_wall_ms": 24434.069832903333, "ttft_ms": 18.778624944388866, "input_usage": {"lookups": 0}, "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": "1,559,015 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "8832bd02543a64471be303a477866133239f43e33b49289138ae98ff9eeb9683"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "530606738f98adb6ab54f509", "attempt_id": "bc07cfc65d59", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:29.469474Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 441.3267499767244, "ttft_ms": 36.082457983866334, "decode_ms": 30.327125103212893}, "total_wall_ms": 24229.96212495491, "ttft_ms": 36.082457983866334, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e2643a9f0bf460edcee39d60", "attempt_id": "dc6949091ea8", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:29.997049Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 281.45662508904934, "ttft_ms": 19.921249942854047, "decode_ms": 266.5066660847515}, "total_wall_ms": 24290.164957987145, "ttft_ms": 19.921249942854047, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d75a6fc9a1f97bd34788c558", "attempt_id": "7f24dc5e80fb", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:30.584856Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 257.5902919052169, "ttft_ms": 44.357583043165505, "decode_ms": 58.302875026129186}, "total_wall_ms": 24082.339750020765, "ttft_ms": 44.357583043165505, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5c0783573a581fa73c1a13a8", "attempt_id": "577ecb805d06", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:30.966201Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 349.04895804356784, "ttft_ms": 120.17429200932384, "decode_ms": 21.23504097107798}, "total_wall_ms": 24212.488417048007, "ttft_ms": 120.17429200932384, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1a765a6573bdcc3a704e4520", "attempt_id": "bda6b7a32dad", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:31.475836Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 449, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23721.36587498244, "hops": {"1": 23721.36587498244}, "receiver_prefill_ms": 270.6541660008952, "ttft_ms": 23.75749999191612, "decode_ms": 35.931042046286166}, "total_wall_ms": 24052.31054208707, "ttft_ms": 23.75749999191612, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6ae587acbf8794149a787ca8", "attempt_id": "6c8abda5d334", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:48.305129Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 435, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 361.4017500076443, "ttft_ms": 38.088415982201695, "decode_ms": 37.329708924517035}, "total_wall_ms": 40636.82916702237, "ttft_ms": 38.088415982201695, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e574621335a01ffc4a86d44f", "attempt_id": "8984baaa1b0c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:48.764911Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 441, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 291.2917499197647, "ttft_ms": 37.644708063453436, "decode_ms": 40.85200000554323}, "total_wall_ms": 40569.79041697923, "ttft_ms": 37.644708063453436, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfb51dd98ada3126dc8c2737", "attempt_id": "c8c57f541aee", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:49.153954Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 439, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 272.1136669861153, "ttft_ms": 54.92408399004489, "decode_ms": 44.97883305884898}, "total_wall_ms": 40572.040500934236, "ttft_ms": 54.92408399004489, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cbb0062c20a684c6c49dba1d", "attempt_id": "8b290d82caef", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:49.545190Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 357.9865420470014, "ttft_ms": 137.11149990558624, "decode_ms": 40.30320805031806}, "total_wall_ms": 40735.46295892447, "ttft_ms": 137.11149990558624, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1f327cffbaea94e8acd4a711", "attempt_id": "c7e9cedf1a2d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:50.099801Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 285.71041603572667, "ttft_ms": 25.10258404072374, "decode_ms": 375.20958308596164}, "total_wall_ms": 40886.10608293675, "ttft_ms": 25.10258404072374, "input_usage": {"lookups": 0}, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8b6e7da907d6efa5d1593c9e", "attempt_id": "058f78b0e1cd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:50.806567Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 257.4868749361485, "ttft_ms": 47.059500007890165, "decode_ms": 592.6133330212906}, "total_wall_ms": 41097.263791947626, "ttft_ms": 47.059500007890165, "input_usage": {"lookups": 0}, "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-79U-1063", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "0de04aed6b20f3db39c86fb65475b79b59798e252a9c06c61105f9cdf5b96aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bcd785061cd6318036b312a1", "attempt_id": "10af0d351194", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:51.725459Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 291.6537920245901, "ttft_ms": 33.506875042803586, "decode_ms": 35.54524993523955}, "total_wall_ms": 40560.790374991484, "ttft_ms": 33.506875042803586, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d5d7f847b1188e8e87d9275b", "attempt_id": "e0a390c4598e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:52.107135Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 436, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 261.318540899083, "ttft_ms": 20.00516711268574, "decode_ms": 35.64604197163135}, "total_wall_ms": 40517.10275001824, "ttft_ms": 20.00516711268574, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2ca3fd7f39284507fda62272", "attempt_id": "887517d3786a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:52.444991Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 451, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 297.90729098021984, "ttft_ms": 56.50791700463742, "decode_ms": 37.06104203592986}, "total_wall_ms": 40591.58225101419, "ttft_ms": 56.50791700463742, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c0f37ba72128faa769c09ec6", "attempt_id": "5be6c4cc8f16", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:52.857422Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 425.45141698792577, "ttft_ms": 27.627917006611824, "decode_ms": 300.981167005375}, "total_wall_ms": 40954.11620894447, "ttft_ms": 27.627917006611824, "input_usage": {"lookups": 0}, "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": "1,559,015", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "10227575e6b3c4a42595de3f6ede24baf18e5e7d76643c8079030af6789a43a5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1a16a9f4d7bca997983b808b", "attempt_id": "d25eb84e2b41", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:53.634594Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 309.9906659917906, "ttft_ms": 19.392999936826527, "decode_ms": 37.36804204527289}, "total_wall_ms": 40566.755583975464, "ttft_ms": 19.392999936826527, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "709ebfcd359370aa2199ba3b", "attempt_id": "caead7671755", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:54.020771Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 320.9559170063585, "ttft_ms": 70.2855410054326, "decode_ms": 404.43712507840246}, "total_wall_ms": 40995.83849997725, "ttft_ms": 70.2855410054326, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dd6c1452e623c793f34a97a6", "attempt_id": "bfbfdd6a933d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:54.837451Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 302.9710000846535, "ttft_ms": 32.15041698422283, "decode_ms": 38.45558303873986}, "total_wall_ms": 40573.66612495389, "ttft_ms": 32.15041698422283, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bd505825679d6ad31de022b6", "attempt_id": "4981668acaac", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:55.231932Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 260.6348330155015, "ttft_ms": 35.15250002965331, "decode_ms": 45.03245803061873}, "total_wall_ms": 40541.111333994195, "ttft_ms": 35.15250002965331, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dedf37d3ba8e25ac44c8a76e", "attempt_id": "28b91354fffd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:20:55.592429Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 449, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40199.46504198015, "hops": {"1": 23721.36587498244, "2": 16478.09916699771}, "receiver_prefill_ms": 275.4081250168383, "ttft_ms": 52.89675004314631, "decode_ms": 53.58808406163007}, "total_wall_ms": 40581.885874969885, "ttft_ms": 52.89675004314631, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a02b60c41b41d0056a64171c", "attempt_id": "0b605b5029ca", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:28.817575Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 435, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 295.7721659913659, "ttft_ms": 39.776916964910924, "decode_ms": 23.898291983641684}, "total_wall_ms": 73381.52195897419, "ttft_ms": 39.776916964910924, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "03d1754075674a67da08495a", "attempt_id": "c46ab74042c8", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:29.197041Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 441, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 287.6692080171779, "ttft_ms": 25.060291052795947, "decode_ms": 37.501041078940034}, "total_wall_ms": 73372.1872511087, "ttft_ms": 25.060291052795947, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e9774379c70a13c503a74f1", "attempt_id": "90c1f45c773d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:29.566160Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 439, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 268.7292500631884, "ttft_ms": 51.564250024966896, "decode_ms": 38.71600003913045}, "total_wall_ms": 73381.01604313124, "ttft_ms": 51.564250024966896, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0927f7ffad5501bbe5204cf1", "attempt_id": "2fcf1b90b842", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:29.944560Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 471.60804097075015, "ttft_ms": 33.03325001616031, "decode_ms": 37.49529202468693}, "total_wall_ms": 73564.23737609293, "ttft_ms": 33.03325001616031, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f88b5b39650b0054f4d449ea", "attempt_id": "32857ed92dee", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:30.507128Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 280.87262506596744, "ttft_ms": 25.908999959938228, "decode_ms": 333.45895796082914}, "total_wall_ms": 73662.24229312502, "ttft_ms": 25.908999959938228, "input_usage": {"lookups": 0}, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "576e0f7cd8304c782d140d76", "attempt_id": "018bbb49dfe7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:31.166379Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 316.34354195557535, "ttft_ms": 66.91154197324067, "decode_ms": 553.978250012733}, "total_wall_ms": 73959.35308502521, "ttft_ms": 66.91154197324067, "input_usage": {"lookups": 0}, "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-79U-1063", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "0de04aed6b20f3db39c86fb65475b79b59798e252a9c06c61105f9cdf5b96aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5476fb0605ff9d5c084fa869", "attempt_id": "2ca4e01c542e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:32.124592Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 281.3850420061499, "ttft_ms": 25.391416042111814, "decode_ms": 37.019249983131886}, "total_wall_ms": 73365.81945908256, "ttft_ms": 25.391416042111814, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2c158e979fdef9327daa07ed", "attempt_id": "79263a682e32", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:32.487881Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 436, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 261.1760830041021, "ttft_ms": 51.421792013570666, "decode_ms": 53.32820909097791}, "total_wall_ms": 73387.95083505102, "ttft_ms": 51.421792013570666, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b7d14cf6152a3677baeb1a31", "attempt_id": "d01b773eea37", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:32.874786Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 451, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 384.3151250621304, "ttft_ms": 119.1998750437051, "decode_ms": 35.21066706161946}, "total_wall_ms": 73560.71779306512, "ttft_ms": 119.1998750437051, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "586b23f1d5a572964da27286", "attempt_id": "43d74cb73381", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:33.434543Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 285.47145798802376, "ttft_ms": 21.426124963909388, "decode_ms": 322.11416598875076}, "total_wall_ms": 73651.0123340413, "ttft_ms": 21.426124963909388, "input_usage": {"lookups": 0}, "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": "1,559,015", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "10227575e6b3c4a42595de3f6ede24baf18e5e7d76643c8079030af6789a43a5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0db1a4700455a7e1725b2920", "attempt_id": "03b698b93192", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:34.082931Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 272.9066669708118, "ttft_ms": 41.75108310300857, "decode_ms": 58.062499971129}, "total_wall_ms": 73394.68700205907, "ttft_ms": 41.75108310300857, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3e46845493b0988e1a030634", "attempt_id": "61658e1eac96", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:34.475978Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 392.66858401242644, "ttft_ms": 86.7149590048939, "decode_ms": 236.74016597215086}, "total_wall_ms": 73738.15579304937, "ttft_ms": 86.7149590048939, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fc2fd70c6d3a5a5157a4a44d", "attempt_id": "e71ec63478b5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:35.212929Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 302.55158396903425, "ttft_ms": 40.92974995728582, "decode_ms": 30.595874995924532}, "total_wall_ms": 73396.16491808556, "ttft_ms": 40.92974995728582, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ffdf87abbb422060beb840f5", "attempt_id": "b11e266eb2e3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:35.609623Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 276.81670896708965, "ttft_ms": 36.941208061762154, "decode_ms": 53.35262499283999}, "total_wall_ms": 73389.53504303936, "ttft_ms": 36.941208061762154, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c1906f0c7eb1f90b69ab0318", "attempt_id": "b05442565a65", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:21:35.998757Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 449, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73021.4465430472, "hops": {"1": 23721.36587498244, "2": 16478.09916699771, "3": 16468.297042069025, "4": 16353.684458998032}, "receiver_prefill_ms": 301.26104201190174, "ttft_ms": 95.23091604933143, "decode_ms": 87.99579204060137}, "total_wall_ms": 73506.87679310795, "ttft_ms": 95.23091604933143, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ceaa027cc97949a395bd2074", "attempt_id": "ea87de58815a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:00.333513Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 383.31704097799957, "ttft_ms": 32.179375062696636, "decode_ms": 31.079957960173488}, "total_wall_ms": 24273.662582971156, "ttft_ms": 32.179375062696636, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9c844e816912a8d9076d0e7d", "attempt_id": "0d2eabdb8576", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:00.800859Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 396.7593749985099, "ttft_ms": 34.07145803794265, "decode_ms": 33.208959037438035}, "total_wall_ms": 24291.040625073947, "ttft_ms": 34.07145803794265, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "12306f3f42ea8fddabbc1883", "attempt_id": "222474b36b4a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:01.284476Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 363.2564999861643, "ttft_ms": 63.45566699746996, "decode_ms": 425.4177500260994}, "total_wall_ms": 24679.310084087774, "ttft_ms": 63.45566699746996, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35.5234", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "005768a0245176a932a31dd800f8ef72b75c295d00e4f081a98f2978ca880367"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5fcc2e3e98a36d82a0c685c1", "attempt_id": "5c756cdefdfa", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:02.157033Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 401.8465420231223, "ttft_ms": 34.58866698201746, "decode_ms": 236.7562080034986}, "total_wall_ms": 24500.34441705793, "ttft_ms": 34.58866698201746, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "76035cadb944b32b9d193ad8", "attempt_id": "24fbbc204a17", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:02.852380Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 410.38291703443974, "ttft_ms": 61.05933408252895, "decode_ms": 462.67362497746944}, "total_wall_ms": 24761.38191705104, "ttft_ms": 61.05933408252895, "input_usage": {"lookups": 0}, "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": "1,844,550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "c3a824f75c26a69ab7e2a0dfde5b194c586a879a105eff635ab136cdb1a4fd6c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d2900a4d4d7ec6ccc6f55e54", "attempt_id": "3b15d7069790", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:03.808872Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 384.3591670738533, "ttft_ms": 38.11233292799443, "decode_ms": 38.332208059728146}, "total_wall_ms": 24287.94362512417, "ttft_ms": 38.11233292799443, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b242e7e126145b1967454fa6", "attempt_id": "af3e11210165", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:04.288933Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 349.6012920513749, "ttft_ms": 36.39537503477186, "decode_ms": 60.457958024926484}, "total_wall_ms": 24273.693709168583, "ttft_ms": 36.39537503477186, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5a619c48228ad7996608ac87", "attempt_id": "1e09925c682e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:04.754899Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 531.482958002016, "ttft_ms": 31.752583920024335, "decode_ms": 37.83579193986952}, "total_wall_ms": 24428.378458018415, "ttft_ms": 31.752583920024335, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1ef6aeb47fc397d7c7867262", "attempt_id": "098706aeb4ba", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:05.377717Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 609, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 385.42545889504254, "ttft_ms": 41.089292033575475, "decode_ms": 39.29725009948015}, "total_wall_ms": 24293.002292048186, "ttft_ms": 41.089292033575475, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a8c0950bdfc32d90e8df12c9", "attempt_id": "c66718f7853e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:05.865462Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 357.67100006341934, "ttft_ms": 37.1997500769794, "decode_ms": 404.21670803334564}, "total_wall_ms": 24626.261625089683, "ttft_ms": 37.1997500769794, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6dc7c43839d913ce2a20f8cf", "attempt_id": "cbc24cf6648e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:06.685363Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 470.94416699837893, "ttft_ms": 39.619875024072826, "decode_ms": 37.19904203899205}, "total_wall_ms": 24374.822084093466, "ttft_ms": 39.619875024072826, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6feed2e207196350a1c5a5f6", "attempt_id": "a4e1b6ffd167", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:07.253951Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 381.9044589763507, "ttft_ms": 39.7102499846369, "decode_ms": 263.67295801173896}, "total_wall_ms": 24512.723626103252, "ttft_ms": 39.7102499846369, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5464.31", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "754be52ae0e9ea0a9d6f318d0a6289c2971ea97dc372d8c4a25cd50e97b97540"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a0fb9013b7824967da59a429", "attempt_id": "7a634cb2c820", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:07.960357Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 558.2975420402363, "ttft_ms": 32.92320796754211, "decode_ms": 35.29862500727177}, "total_wall_ms": 24453.75450013671, "ttft_ms": 32.92320796754211, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a3e455a7975832c0cf45e2b", "attempt_id": "f114152f2ec4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:08.609707Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 381.6480420064181, "ttft_ms": 23.63195805810392, "decode_ms": 38.91416697297245}, "total_wall_ms": 24271.478834096342, "ttft_ms": 23.63195805810392, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c4468f15a89e9a57965012fa", "attempt_id": "91d9f9c45883", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:09.074340Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1172, "cache_tensor_bytes": null, "serialized_payload_bytes": 1172, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23826.40841708053, "hops": {"1": 23826.40841708053}, "receiver_prefill_ms": 378.4869590308517, "ttft_ms": 19.468333921395242, "decode_ms": 55.42612506542355}, "total_wall_ms": 24280.736792017706, "ttft_ms": 19.468333921395242, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "6c4584da6785b21b7a948215b5067865f5b329d9973723d1578fbb55483e4d52", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5f3b26e2a5b7e22b1223892d", "attempt_id": "d6a57ad1bb9a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:32.403786Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 386.4008749369532, "ttft_ms": 40.27341597247869, "decode_ms": 24.512375006452203}, "total_wall_ms": 47131.31566706579, "ttft_ms": 40.27341597247869, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "40258cbb1718088d20ece4e8", "attempt_id": "45c9ee9be7f6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:32.875092Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 372.0741670113057, "ttft_ms": 23.460375028662384, "decode_ms": 72.34891701955348}, "total_wall_ms": 47147.8880001232, "ttft_ms": 23.460375028662384, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c9bc2c12e95955956f26e11", "attempt_id": "bff77e5f3195", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:33.363165Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 373.73212503734976, "ttft_ms": 91.67275007348508, "decode_ms": 345.14108300209045}, "total_wall_ms": 47490.7852911856, "ttft_ms": 91.67275007348508, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "35.5234", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "005768a0245176a932a31dd800f8ef72b75c295d00e4f081a98f2978ca880367"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62845dd5d9b9e1d34de6cb92", "attempt_id": "96ae2c1b5c11", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:34.197115Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 402.485708007589, "ttft_ms": 26.23512491118163, "decode_ms": 196.5241669677198}, "total_wall_ms": 47305.26679207105, "ttft_ms": 26.23512491118163, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62cc47cc834d856d1ad9ed4b", "attempt_id": "1bf2c38e3198", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:34.842357Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 377.29199999012053, "ttft_ms": 55.17533398233354, "decode_ms": 426.80958309210837}, "total_wall_ms": 47539.366041193716, "ttft_ms": 55.17533398233354, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4b11e5861e200bf0311258d4", "attempt_id": "2bb549e0c591", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:35.723108Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 386.1412910046056, "ttft_ms": 39.2737080110237, "decode_ms": 26.133332983590662}, "total_wall_ms": 47131.5720831044, "ttft_ms": 39.2737080110237, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ceb5acf5bd95334a50ee4d7e", "attempt_id": "904577170d95", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:36.193969Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 338.3196670329198, "ttft_ms": 36.56287491321564, "decode_ms": 52.72541707381606}, "total_wall_ms": 47107.855500187725, "ttft_ms": 36.56287491321564, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cc27b3fa01028556a7e33ef1", "attempt_id": "6932efd32365", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:36.643406Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 445.64329204149544, "ttft_ms": 135.50695800222456, "decode_ms": 40.59983300976455}, "total_wall_ms": 47301.67979118414, "ttft_ms": 135.50695800222456, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "df12b8cd48c5cc0deb38c6e4", "attempt_id": "0e297a6d132e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:37.284479Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 609, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 396.07466699089855, "ttft_ms": 35.74354201555252, "decode_ms": 39.25804200116545}, "total_wall_ms": 47151.15279215388, "ttft_ms": 35.74354201555252, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "01cf0b1a0322c9b5a1057510", "attempt_id": "b6da093cb966", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:37.775677Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 388.79774999804795, "ttft_ms": 24.10387503914535, "decode_ms": 350.04091705195606}, "total_wall_ms": 47443.011583178304, "ttft_ms": 24.10387503914535, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "63e3043dd702230e6267d9ce", "attempt_id": "c7f3908275c5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:38.559449Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 611, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 524.1479580290616, "ttft_ms": 41.39229201246053, "decode_ms": 43.39424998033792}, "total_wall_ms": 47288.90208317898, "ttft_ms": 41.39229201246053, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5f6278d299420146f8ba5a0a", "attempt_id": "6abf01ff9350", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:39.189201Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 398.1404999503866, "ttft_ms": 24.93908302858472, "decode_ms": 258.1785829970613}, "total_wall_ms": 47361.19533411693, "ttft_ms": 24.93908302858472, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4663.36", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "a07b874ce4e807caaa33f72eb401eedac860ff6ec404621a3ec4be87e3685d87"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1d731f610bd2ab8aee700148", "attempt_id": "aad61bdc1570", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:39.890827Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 366.22100009117275, "ttft_ms": 120.91095792129636, "decode_ms": 99.89983297418803}, "total_wall_ms": 47266.99962513521, "ttft_ms": 120.91095792129636, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a8b6d7a70fb3cbf6d46025f8", "attempt_id": "b5ef073aa93e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:40.499099Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 376.2676250189543, "ttft_ms": 31.76366607658565, "decode_ms": 30.372792039997876}, "total_wall_ms": 47118.30195912626, "ttft_ms": 31.76366607658565, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6eb224fcdc77362a83b7f3d1", "attempt_id": "8c8f556a11ef", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:22:40.958386Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1161, "cache_tensor_bytes": null, "serialized_payload_bytes": 1161, "side_information_bytes": 0, "fetched_bytes": 836, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46679.23775012605, "hops": {"1": 23826.40841708053, "2": 22852.82933304552}, "receiver_prefill_ms": 393.37129099294543, "ttft_ms": 33.50308409426361, "decode_ms": 34.41900003235787}, "total_wall_ms": 47141.36466605123, "ttft_ms": 33.50308409426361, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ef10d21f73ab5677e6eb35d30c0602c132dabdab8f7ac38eec07514083f21010", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bae6428628070692d3ae3340", "attempt_id": "2fc3aedb67d4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:20.323834Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 509, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 311.13016698509455, "ttft_ms": 42.49900009017438, "decode_ms": 19.375500036403537}, "total_wall_ms": 85935.15108304564, "ttft_ms": 42.49900009017438, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "70d4af435cdce2147717c834", "attempt_id": "dc4ac7a7761f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:20.716383Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 515, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 334.147582994774, "ttft_ms": 60.100542032159865, "decode_ms": 59.14783396292478}, "total_wall_ms": 86015.44445811305, "ttft_ms": 60.100542032159865, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d79d92a57d52ac291ad0045f", "attempt_id": "a4b48e432f0e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:21.188724Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 513, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 471.825624932535, "ttft_ms": 32.00954105705023, "decode_ms": 222.15262497775257}, "total_wall_ms": 86288.15937519539, "ttft_ms": 32.00954105705023, "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": "9.1829", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "7be954dd35132848588fff1074a4e446b3736859874b66427ec9139235913505"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ed001c81a4da44c1ccad1d1c", "attempt_id": "c952c81973ff", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:21.935737Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 511, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 296.3970829732716, "ttft_ms": 30.47183307353407, "decode_ms": 252.19795806333423}, "total_wall_ms": 86141.1111241905, "ttft_ms": 30.47183307353407, "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": "Rasheed Mäkelä", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "e35e1a6499b45eff74141ed4819137b73a904015adfe78e6fa03b5194d802836"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1f4605c3525c7ea0ba01a5fc", "attempt_id": "7d4c626f4b91", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:22.534335Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 517, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 485.3400000138208, "ttft_ms": 81.2257919460535, "decode_ms": 276.149042067118}, "total_wall_ms": 86404.83912511263, "ttft_ms": 81.2257919460535, "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": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "66e8af6d1e03f35c94bfb2fe", "attempt_id": "6a9bec144938", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:23.398391Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 520, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 367.8697500145063, "ttft_ms": 33.72379101347178, "decode_ms": 34.883041982539}, "total_wall_ms": 85998.55008313898, "ttft_ms": 33.72379101347178, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0bb8e4f9c4786f0e2465da0a", "attempt_id": "52957607829d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:23.855526Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 511, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 327.45270791929215, "ttft_ms": 38.02479198202491, "decode_ms": 46.71120794955641}, "total_wall_ms": 85974.24787411, "ttft_ms": 38.02479198202491, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "45c1a65491f8d4c969b37869", "attempt_id": "990fbd579050", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:24.288681Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 510, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 465.1638329960406, "ttft_ms": 21.980290999636054, "decode_ms": 34.27454095799476}, "total_wall_ms": 86083.59716611449, "ttft_ms": 21.980290999636054, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b5c1449290c3fcb42ab33ca0", "attempt_id": "78b5c6bad587", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:24.831253Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 525, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 386.09254197217524, "ttft_ms": 26.262625004164875, "decode_ms": 34.96362501755357}, "total_wall_ms": 86009.322833037, "ttft_ms": 26.262625004164875, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b604abaa11120e518064d997", "attempt_id": "b333bc616a76", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:25.299051Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 517, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 348.9590419922024, "ttft_ms": 49.635584000498056, "decode_ms": 510.96441596746445}, "total_wall_ms": 86471.66991617996, "ttft_ms": 49.635584000498056, "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": "2968560", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "71eefa6bcfef23ffcab83a41e37fc18601749a14cc43068c167f2b87d81c4fc1"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "84cf28bdf18be287051dce4b", "attempt_id": "7f22609c5e75", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:26.227509Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 527, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 379.32599999476224, "ttft_ms": 36.612584022805095, "decode_ms": 28.256250079721212}, "total_wall_ms": 86006.31366611924, "ttft_ms": 36.612584022805095, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "35c466995b43154e7550b67d", "attempt_id": "32bbc0445eb4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:26.690771Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 521, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 371.93174997810274, "ttft_ms": 20.88245889171958, "decode_ms": 352.89220907725394}, "total_wall_ms": 86307.79716616962, "ttft_ms": 20.88245889171958, "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": "4663.36", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "a07b874ce4e807caaa33f72eb401eedac860ff6ec404621a3ec4be87e3685d87"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2c3d025f0c91592d580363cd", "attempt_id": "3ddb5d1ede77", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:27.455825Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 518, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 495.2845829539001, "ttft_ms": 43.38091693352908, "decode_ms": 40.062166983261704}, "total_wall_ms": 86140.9633330768, "ttft_ms": 43.38091693352908, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62b322c2b0d475e409168596", "attempt_id": "2ae27a2c07f5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:28.056211Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 521, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 362.3665829654783, "ttft_ms": 39.301375043578446, "decode_ms": 36.792082944884896}, "total_wall_ms": 86000.54654211272, "ttft_ms": 39.301375043578446, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28bc00e52f289dd79aad0c01", "attempt_id": "c1164f381cd8", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:28.516468Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 523, "message_tokens": 428, "message_utf8_bytes": 976, "cache_tensor_bytes": null, "serialized_payload_bytes": 976, "side_information_bytes": 0, "fetched_bytes": 2508, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 85561.43533310387, "hops": {"1": 23826.40841708053, "2": 22852.82933304552, "3": 19478.87875000015, "4": 19403.318832977675}, "receiver_prefill_ms": 335.33791603986174, "ttft_ms": 43.78662502858788, "decode_ms": 59.35533298179507}, "total_wall_ms": 86000.67150010727, "ttft_ms": 43.78662502858788, "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": {"ledger": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 27, "retention_rate": 0.40298507462686567, "retention_by_kind": {"person": 0.5, "date": 0.1, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1172}, "2": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1161}, "3": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 977}, "4": {"facts_total": 67, "facts_retained": 26, "retention_rate": 0.3880597014925373, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.4666666666666667, "status": 0.75, "place": 0.0, "identifier": 0.4166666666666667, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 976}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "ce1cb970112e3d7971298f90d1afefdbf3bee4ebf28f75b34f1fabcdb79f5115", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "102959cc4d190d3e3dd21d05", "attempt_id": "3be19b23734c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:52.357118Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 435, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 286.7255830205977, "ttft_ms": 36.98245796840638, "decode_ms": 23.31654098816216}, "total_wall_ms": 23727.558749960735, "ttft_ms": 36.98245796840638, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f8bc5a470eb5ae31fe2d12f5", "attempt_id": "0941ddba6034", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:52.724896Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 441, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 268.44879204872996, "ttft_ms": 54.55649993382394, "decode_ms": 60.13104203157127}, "total_wall_ms": 23763.865708955564, "ttft_ms": 54.55649993382394, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e4fca67ff9cfa8c67662e40c", "attempt_id": "4c0e1d1b71b3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:53.130700Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 439, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 444.94220800697803, "ttft_ms": 26.537166093476117, "decode_ms": 37.790499976836145}, "total_wall_ms": 23889.827791950665, "ttft_ms": 26.537166093476117, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "95f2bfbfb27e66b61a5393e6", "attempt_id": "e105677fe7a2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:53.659063Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 291.33333300706, "ttft_ms": 36.30550007801503, "decode_ms": 30.023292056284845}, "total_wall_ms": 23738.31174999941, "ttft_ms": 36.30550007801503, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb7fa6bf19fddfb70cca339a", "attempt_id": "4becaf3d97d5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:54.036926Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 286.7097499547526, "ttft_ms": 23.230790975503623, "decode_ms": 617.4542920198292}, "total_wall_ms": 24307.933083968237, "ttft_ms": 23.230790975503623, "input_usage": {"lookups": 0}, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "87598b19b2bcd483addbab97", "attempt_id": "71d4c73b4d0a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:54.983255Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 299.07825007103384, "ttft_ms": 39.23849994316697, "decode_ms": 351.6769160050899}, "total_wall_ms": 24070.569291943684, "ttft_ms": 39.23849994316697, "input_usage": {"lookups": 0}, "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-79U-1063", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "0de04aed6b20f3db39c86fb65475b79b59798e252a9c06c61105f9cdf5b96aef"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a3b7b8569ab47183e3dca8d8", "attempt_id": "f49f1ffd07c9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:55.693313Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 437, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 257.3639589827508, "ttft_ms": 55.836457991972566, "decode_ms": 58.456750004552305}, "total_wall_ms": 23752.341457875445, "ttft_ms": 55.836457991972566, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a2841e237427e52a158abaf7", "attempt_id": "e6136c4aec43", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:56.085333Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 436, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 425.1930830068886, "ttft_ms": 38.402333040721714, "decode_ms": 23.817250039428473}, "total_wall_ms": 23867.98674997408, "ttft_ms": 38.402333040721714, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bd172d3073a3a05bd43742f2", "attempt_id": "881fa7e8af7e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:56.592735Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 451, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 313.5162080870941, "ttft_ms": 37.62179100885987, "decode_ms": 22.72883302066475}, "total_wall_ms": 23754.49379195925, "ttft_ms": 37.62179100885987, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebc563926be88022bd6eed11", "attempt_id": "70df3a4ae13e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:56.985832Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 443, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 289.6867080125958, "ttft_ms": 36.89400001894683, "decode_ms": 557.3245419654995}, "total_wall_ms": 24264.45429190062, "ttft_ms": 36.89400001894683, "input_usage": {"lookups": 0}, "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": "1,559,015 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "8832bd02543a64471be303a477866133239f43e33b49289138ae98ff9eeb9683"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "361c3bb237efaa558690a59e", "attempt_id": "6b77ff3079a8", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:57.888610Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 357.4866669951007, "ttft_ms": 25.435500079765916, "decode_ms": 45.39937502704561}, "total_wall_ms": 23809.027957962826, "ttft_ms": 25.435500079765916, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d8b60319f401c0c6e339eb99", "attempt_id": "1046dd10d10e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:58.336892Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 285.0591669557616, "ttft_ms": 41.55999992508441, "decode_ms": 210.0118329981342}, "total_wall_ms": 23917.4062499078, "ttft_ms": 41.55999992508441, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f6a48438961e069f465254e5", "attempt_id": "cae1ecec7c21", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:58.894586Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 308.74970904551446, "ttft_ms": 55.80912495497614, "decode_ms": 94.46404094342142}, "total_wall_ms": 23839.73695896566, "ttft_ms": 55.80912495497614, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8e2cef1b02523f1b698ada85", "attempt_id": "9d1ed8ced988", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:59.374138Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 447, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 368.63787507172674, "ttft_ms": 38.52420800831169, "decode_ms": 38.26316702179611}, "total_wall_ms": 23826.059416984208, "ttft_ms": 38.52420800831169, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3b22ded292e42c583fa4f1b1", "attempt_id": "91754289fbad", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:23:59.841292Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 449, "message_tokens": 353, "message_utf8_bytes": 829, "cache_tensor_bytes": null, "serialized_payload_bytes": 829, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23379.991541965865, "hops": {"1": 23379.991541965865}, "receiver_prefill_ms": 278.5208340501413, "ttft_ms": 40.0078339735046, "decode_ms": 26.421791058965027}, "total_wall_ms": 23725.564208929427, "ttft_ms": 40.0078339735046, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bf1c7d5c36c8e7c5baf81b3a1a2b3ce50b7f0406ed34cf69978346c3fad620ea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ee56d3f0524c4c10d203538a", "attempt_id": "5e8cdf738cef", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:16.738193Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 302.9905828880146, "ttft_ms": 44.648417038843036, "decode_ms": 24.544375017285347}, "total_wall_ms": 40284.4987079734, "ttft_ms": 44.648417038843036, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "352edd0161cdc7b5562bf25b", "attempt_id": "99e9ccada2a9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:17.131611Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 450, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 303.4886250970885, "ttft_ms": 37.49220794998109, "decode_ms": 25.792208034545183}, "total_wall_ms": 40278.96404196508, "ttft_ms": 37.49220794998109, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "387f5e6921207d3cb7be83f6", "attempt_id": "d57281f0c8ad", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:17.517520Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 448, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 289.76804099511355, "ttft_ms": 36.50816704612225, "decode_ms": 55.6773750577122}, "total_wall_ms": 40294.45429192856, "ttft_ms": 36.50816704612225, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "757515985960de68919cc65d", "attempt_id": "d018bc4d7533", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:17.920000Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 445.9719170117751, "ttft_ms": 40.94216600060463, "decode_ms": 30.049749999307096}, "total_wall_ms": 40429.20458386652, "ttft_ms": 40.94216600060463, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2e2bbf416e172b94264d3a2c", "attempt_id": "677ae4fd15f7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:18.456455Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 452, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 300.75991700869054, "ttft_ms": 26.087583042681217, "decode_ms": 318.9650409622118}, "total_wall_ms": 40558.378124958836, "ttft_ms": 26.087583042681217, "input_usage": {"lookups": 1}, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6071a700acd2464e23f61122", "attempt_id": "6291a21c21c1", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:19.122568Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 455, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 344.870708999224, "ttft_ms": 59.02187491301447, "decode_ms": 98.92279095947742}, "total_wall_ms": 40415.150000015274, "ttft_ms": 59.02187491301447, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e41f76103473da9e4a8b8d20", "attempt_id": "0caa0e83138b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:19.647733Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 364.4460419891402, "ttft_ms": 39.320250041782856, "decode_ms": 33.73087500222027}, "total_wall_ms": 40349.79195892811, "ttft_ms": 39.320250041782856, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e8738db1c6ed5989f2e4ae36", "attempt_id": "34066abe5b18", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:20.108560Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 445, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 283.99212507065386, "ttft_ms": 40.50329199526459, "decode_ms": 29.514542082324624}, "total_wall_ms": 40266.25233294908, "ttft_ms": 40.50329199526459, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aa2fac00c9ec243229c1e932", "attempt_id": "5b079b699600", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:20.482373Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 460, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 297.86554200109094, "ttft_ms": 45.49179202876985, "decode_ms": 57.97366693150252}, "total_wall_ms": 40313.718333956786, "ttft_ms": 45.49179202876985, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd58e7c856f777d79453840f", "attempt_id": "a07d67c6318c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:20.903397Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 452, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 301.4050000347197, "ttft_ms": 91.00808296352625, "decode_ms": 455.15141705982387}, "total_wall_ms": 40759.96699987445, "ttft_ms": 91.00808296352625, "input_usage": {"lookups": 1}, "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": "1,559,015 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "8832bd02543a64471be303a477866133239f43e33b49289138ae98ff9eeb9683"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "da6a8aa580b75bb892950012", "attempt_id": "b8f3b584487a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:21.772907Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 462, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 320.92708302661777, "ttft_ms": 41.46316694095731, "decode_ms": 32.18516591005027}, "total_wall_ms": 40306.91829195712, "ttft_ms": 41.46316694095731, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d08bb3aadbd52d8ab0846093", "attempt_id": "e1a030c79842", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:22.188443Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 456, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 269.42766609136015, "ttft_ms": 55.151958018541336, "decode_ms": 486.7570420028642}, "total_wall_ms": 40723.60149992164, "ttft_ms": 55.151958018541336, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e258b005ba5e3d1f548a6e4e", "attempt_id": "a9a819034d75", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:23.019119Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 323.0228750035167, "ttft_ms": 37.54879196640104, "decode_ms": 24.352291016839445}, "total_wall_ms": 40297.18762496486, "ttft_ms": 37.54879196640104, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c72db7402867660dcb15bb54", "attempt_id": "b260619a84e9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:23.425093Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 456, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 303.7457090103999, "ttft_ms": 25.633750017732382, "decode_ms": 36.20162501465529}, "total_wall_ms": 40277.752541936934, "ttft_ms": 25.633750017732382, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "274251727efaa04fd34f5c2a", "attempt_id": "cb78694690b6", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:23.810878Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 458, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 593, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39911.63474996574, "hops": {"1": 23379.991541965865, "2": 16531.643207999878}, "receiver_prefill_ms": 284.06133397947997, "ttft_ms": 36.66966699529439, "decode_ms": 63.438915996812284}, "total_wall_ms": 40296.373041928746, "ttft_ms": 36.66966699529439, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4001862640cb0a5df6d5f766", "attempt_id": "727aa65da2c0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:57.347290Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 444, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 301.26854206901044, "ttft_ms": 42.17458295170218, "decode_ms": 24.67158401850611}, "total_wall_ms": 73408.9891249314, "ttft_ms": 42.17458295170218, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "064d02d31a49c0ce916e37e8", "attempt_id": "d9a7fbd131a0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:57.735672Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 450, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 304.5021661091596, "ttft_ms": 20.19075001589954, "decode_ms": 20.439250045455992}, "total_wall_ms": 73385.93195891008, "ttft_ms": 20.19075001589954, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9db5089e8984f87ba3b9f144", "attempt_id": "c9be01aa258a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:58.100622Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 448, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 318.1165410205722, "ttft_ms": 59.11674990784377, "decode_ms": 102.33541601337492}, "total_wall_ms": 73520.59833391104, "ttft_ms": 59.11674990784377, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f896855002bd940d2e69d268", "attempt_id": "750fe10ea27b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:58.602257Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 391.698915977031, "ttft_ms": 39.86470808740705, "decode_ms": 40.56512494571507}, "total_wall_ms": 73512.92391796596, "ttft_ms": 39.86470808740705, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4594c42fef8adc3d3dbf582b", "attempt_id": "96836af2b3c7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:59.096829Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 452, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 321.59191695973277, "ttft_ms": 41.58612503670156, "decode_ms": 322.06549996044487}, "total_wall_ms": 73726.2132918695, "ttft_ms": 41.58612503670156, "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": "1,844,550 CAD", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 0.6666666666666666}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "f18fbfeb541721d80ae58d07ce145913b9176c27891ed146a5768d81c1484556"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a4193bc9666db32c78e50ec0", "attempt_id": "e24bde0215de", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:24:59.803102Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 455, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 309.24633296672255, "ttft_ms": 91.43220796249807, "decode_ms": 134.09495807718486}, "total_wall_ms": 73575.70495898835, "ttft_ms": 91.43220796249807, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f04d8a6be60f3f462a2fcc15", "attempt_id": "77881f931825", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:00.359568Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 446, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 275.2304580062628, "ttft_ms": 39.15229090489447, "decode_ms": 31.648250063881278}, "total_wall_ms": 73387.06058391836, "ttft_ms": 39.15229090489447, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "018b3266d18ba7d025327329", "attempt_id": "da351f4a6064", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:00.725507Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 445, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 282.65054197981954, "ttft_ms": 38.00283302552998, "decode_ms": 38.594416924752295}, "total_wall_ms": 73400.4199168412, "ttft_ms": 38.00283302552998, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9b1e9eb1fbd1b1a6fd0755fa", "attempt_id": "697bd42875ec", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:01.107734Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 460, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 302.050375030376, "ttft_ms": 43.58024999964982, "decode_ms": 61.66599993593991}, "total_wall_ms": 73448.16404290032, "ttft_ms": 43.58024999964982, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2c6ee878b42e22501bdf7701", "attempt_id": "a590051679f5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:01.536573Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 452, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 356.70591704547405, "ttft_ms": 118.32308303564787, "decode_ms": 306.379416026175}, "total_wall_ms": 73822.29791791178, "ttft_ms": 118.32308303564787, "input_usage": {"lookups": 3}, "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": "1,559,015", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "10227575e6b3c4a42595de3f6ede24baf18e5e7d76643c8079030af6789a43a5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c653b8906575bdedd338e11", "attempt_id": "9bd2990b763d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:02.337727Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 462, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 311.60412495955825, "ttft_ms": 25.082250009290874, "decode_ms": 36.65304207243025}, "total_wall_ms": 73414.12566800136, "ttft_ms": 25.082250009290874, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "59fa904834c716c9be1b4258", "attempt_id": "a7484f8f39dc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:02.731268Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 456, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 322.5153749808669, "ttft_ms": 65.01204206142575, "decode_ms": 491.6987500619143}, "total_wall_ms": 73920.1339998981, "ttft_ms": 65.01204206142575, "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": "1844550", "evidence_artifact": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "23de9db5e2c3413d7142a9e77e743ed168ec6de65abf6413935059d107bb8c17"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4239884db57cedbbe9761ff3", "attempt_id": "7f2108093769", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:03.631265Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 453, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 324.3187499465421, "ttft_ms": 38.32495806273073, "decode_ms": 30.187000054866076}, "total_wall_ms": 73433.79383394495, "ttft_ms": 38.32495806273073, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "275b1610883ef36b980b02be", "attempt_id": "283bf2e63b39", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:04.045383Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 456, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 303.3868749625981, "ttft_ms": 21.12766692880541, "decode_ms": 18.44379201065749}, "total_wall_ms": 73383.74166795984, "ttft_ms": 21.12766692880541, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bd3d7a438a85933a4925d4ad", "attempt_id": "b35e0e2de113", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:04.408210Z", "task_family": "synthetic", "document_id": "syn_02003_2000", "document_hash": "c88d189ebb79e2f792c48ae6111a7e7bf78acb1c3c41f7ea40312f6e47e3dbd8", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02003_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 1989, "query_tokens": 458, "message_tokens": 363, "message_utf8_bytes": 860, "cache_tensor_bytes": null, "serialized_payload_bytes": 860, "side_information_bytes": 0, "fetched_bytes": 1779, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73040.23908392992, "hops": {"1": 23379.991541965865, "2": 16531.643207999878, "3": 16547.03524999786, "4": 16581.569083966315}, "receiver_prefill_ms": 316.0562500124797, "ttft_ms": 47.386125079356134, "decode_ms": 59.93462505284697}, "total_wall_ms": 73464.37804191373, "ttft_ms": 47.386125079356134, "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": {"ledger": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "ledger_by_hop": {"1": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 829}, "2": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "3": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}, "4": {"facts_total": 67, "facts_retained": 16, "retention_rate": 0.23880597014925373, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1, "amount": 0.2, "status": 0.75, "place": 0.0, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 860}}}, "reference_key_id": "keys/synthetic/syn_02003_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/E4_pilot.log", "payload_hash": "bb3bff7982616ec5d81fe45b1fd8ec1cb9ad1cf66ca54ef9d5736e0d2d8991cf", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b7d792d58c2766128e0d9d6f", "attempt_id": "f14220f39f2a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:31.783916Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 372.17479094397277, "ttft_ms": 24.27283301949501, "decode_ms": 59.19683398678899}, "total_wall_ms": 27309.2247081222, "ttft_ms": 24.27283301949501, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bff4da5b96a85a87a127afc9", "attempt_id": "527cefd2e623", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:32.263686Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 438.99045896250755, "ttft_ms": 94.19450000859797, "decode_ms": 97.88216603919864}, "total_wall_ms": 27484.6122500021, "ttft_ms": 94.19450000859797, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4fde9d8cfae161e8afee5233", "attempt_id": "e0ae184e4c72", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:32.920367Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 598, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 401.53212507721037, "ttft_ms": 27.07879210356623, "decode_ms": 235.07262498605996}, "total_wall_ms": 27517.322833067738, "ttft_ms": 27.07879210356623, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "839a2a6242ebe514155c8829", "attempt_id": "a96fd01876be", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:33.607834Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 350.9956659981981, "ttft_ms": 60.02820807043463, "decode_ms": 55.47616700641811}, "total_wall_ms": 27320.31724997796, "ttft_ms": 60.02820807043463, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5be228350dd8068492377fbe", "attempt_id": "6a1a6b8c45e0", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:34.098899Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 545.8219579886645, "ttft_ms": 41.7868330841884, "decode_ms": 24.440958979539573}, "total_wall_ms": 27465.46337509062, "ttft_ms": 41.7868330841884, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e75ece2c2962f2892b280bfd", "attempt_id": "1d85f011ed2a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:34.734983Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 388.6254170211032, "ttft_ms": 32.828583964146674, "decode_ms": 31.636375002563}, "total_wall_ms": 27306.508833076805, "ttft_ms": 32.828583964146674, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c0644ec9bd9a717fe5f9c1dd", "attempt_id": "792f4f7ffe4c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:35.210797Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 354.618625016883, "ttft_ms": 35.82249989267439, "decode_ms": 61.322416993789375}, "total_wall_ms": 27305.417541065253, "ttft_ms": 35.82249989267439, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aac7339e0797a50e40af15c2", "attempt_id": "fded988c86f7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:35.686216Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 427.19383304938674, "ttft_ms": 129.50841698329896, "decode_ms": 40.18591693602502}, "total_wall_ms": 27450.380792142823, "ttft_ms": 129.50841698329896, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "yes", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "8a798890fe93817163b10b5f7bd2ca4d25d84c52739a645a889c173eee7d9d3d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "33b68411d0939a094803c1e0", "attempt_id": "e35128f1ab18", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:36.306692Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 388.8090840773657, "ttft_ms": 39.680875022895634, "decode_ms": 149.20012489892542}, "total_wall_ms": 27431.17200001143, "ttft_ms": 39.680875022895634, "input_usage": {"lookups": 0}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "452e6825d23ca436f873089a", "attempt_id": "2774c1e5a007", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:36.909021Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 350.53233394864947, "ttft_ms": 43.21204195730388, "decode_ms": 52.66233300790191}, "total_wall_ms": 27299.890125053935, "ttft_ms": 43.21204195730388, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7790942933bd7cd7dd808d69", "attempt_id": "857a4655de01", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:37.379081Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 615, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 565.0040420005098, "ttft_ms": 37.51512500457466, "decode_ms": 277.384499902837}, "total_wall_ms": 27733.374083065428, "ttft_ms": 37.51512500457466, "input_usage": {"lookups": 0}, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "82ff893c7e594391df7ff4bb", "attempt_id": "6af09e242b90", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:38.283183Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 379.3443749891594, "ttft_ms": 21.886167000047863, "decode_ms": 1709.278916940093}, "total_wall_ms": 28963.989833020605, "ttft_ms": 21.886167000047863, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 38}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "29.7+24.68+20.7+13.3+21.05+1.17=111.20", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "22eda8802dc2949095cea9119990e1f585a51a909516d797763da8b0c1a15061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "58b5a982b66ee47784e4f238", "attempt_id": "534960df0219", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:40.417021Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 546.6458749724552, "ttft_ms": 23.56212493032217, "decode_ms": 134.36087500303984}, "total_wall_ms": 27558.09470906388, "ttft_ms": 23.56212493032217, "input_usage": {"lookups": 0}, "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": "1.17", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "3d7a51acc5cb98f4e350eed557c0c4a7034a1fdb057a8fc2b38cf2138980b74b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62a8566c47dcdc61279c2145", "attempt_id": "339a32275d1d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:41.147401Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 378.12295893672854, "ttft_ms": 31.021917006000876, "decode_ms": 29.301790986210108}, "total_wall_ms": 27291.806333116256, "ttft_ms": 31.021917006000876, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4e1f6f0327615fa69827aa4e", "attempt_id": "ed6d24579196", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:25:41.608582Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26852.64925006777, "hops": {"1": 26852.64925006777}, "receiver_prefill_ms": 351.7489160876721, "ttft_ms": 44.14537502452731, "decode_ms": 51.70558299869299}, "total_wall_ms": 27301.027750014327, "ttft_ms": 44.14537502452731, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "618191ae06d58bfb582e2dd8", "attempt_id": "b87b01818b45", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:00.721388Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 496, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 331.0272920643911, "ttft_ms": 33.82449992932379, "decode_ms": 33.60695799347013}, "total_wall_ms": 45889.89329105243, "ttft_ms": 33.82449992932379, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e23509e0de9363ca768d8aeb", "attempt_id": "56c949cc76bd", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:01.145448Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 504, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 311.2558330176398, "ttft_ms": 21.39741601422429, "decode_ms": 27.12566696573049}, "total_wall_ms": 45851.28316609189, "ttft_ms": 21.39741601422429, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a9c7272d1c6363ca7a0fc8dc", "attempt_id": "feb387d205f3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:01.531159Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 500, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 295.2322500059381, "ttft_ms": 59.34516608249396, "decode_ms": 427.7937499573454}, "total_wall_ms": 46273.755665984936, "ttft_ms": 59.34516608249396, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "70821b4708237fbe6a4447ee", "attempt_id": "b7798277174d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:02.340759Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 499, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 311.4418330369517, "ttft_ms": 35.17941595055163, "decode_ms": 30.651208944618702}, "total_wall_ms": 45868.685458088294, "ttft_ms": 35.17941595055163, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "082354c27adbfa2e258b6d95", "attempt_id": "f820146e5696", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:02.743793Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 507, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 306.6594999982044, "ttft_ms": 21.316332975402474, "decode_ms": 18.217167002148926}, "total_wall_ms": 45837.5134159578, "ttft_ms": 21.316332975402474, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "971c12b0c016d660100499f1", "attempt_id": "7d0aa5cab0ba", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:03.114549Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 506, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 308.9340419974178, "ttft_ms": 59.488374972715974, "decode_ms": 58.23504098225385}, "total_wall_ms": 45918.23287401348, "ttft_ms": 59.488374972715974, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1c897c9c792f82e55eaee061", "attempt_id": "e6914d91adc1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:03.567415Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 495, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 457.017624983564, "ttft_ms": 35.68529093172401, "decode_ms": 23.91662506852299}, "total_wall_ms": 46007.97291600611, "ttft_ms": 35.68529093172401, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "536a533cd403069d7c939299", "attempt_id": "37ce42406a42", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:04.108242Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 499, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 307.69291694741696, "ttft_ms": 37.67183294985443, "decode_ms": 24.110457976348698}, "total_wall_ms": 45860.73199904058, "ttft_ms": 37.67183294985443, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5a49f6b9fb331b45ed96295e", "attempt_id": "8726a6697030", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:04.501324Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 510, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 287.3552080709487, "ttft_ms": 50.450041075237095, "decode_ms": 231.50495905429125}, "total_wall_ms": 46061.116374097764, "ttft_ms": 50.450041075237095, "input_usage": {"lookups": 0}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "034b94b380a325422937aefb", "attempt_id": "247aac317ad7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:05.095640Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 505, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 487.8680830588564, "ttft_ms": 31.454583047889173, "decode_ms": 42.631874908693135}, "total_wall_ms": 46053.324582986534, "ttft_ms": 31.454583047889173, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "79bd5de6db4f5b134e13dc02", "attempt_id": "e36c955cafa2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:05.683517Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 517, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 361.04458407498896, "ttft_ms": 28.20029202848673, "decode_ms": 280.4851250257343}, "total_wall_ms": 46161.35408205446, "ttft_ms": 28.20029202848673, "input_usage": {"lookups": 0}, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "33398bc5c7ba04e890b0c48d", "attempt_id": "9efe47d9dd94", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:06.377431Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 507, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 341.7964590480551, "ttft_ms": 98.99083292111754, "decode_ms": 825.6729580461979}, "total_wall_ms": 46757.870999048464, "ttft_ms": 98.99083292111754, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 21}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "29.7+24.68+20.7=75.08", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "20ee42c223ccaad7a163f5f49cedc190fe17e1e3328a9903f5c6643f332c7be7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "35c9d822d07caf1c72ae8c83", "attempt_id": "7d696d2daca3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:07.667664Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 504, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 301.4084589667618, "ttft_ms": 60.15545898117125, "decode_ms": 347.32687496580184}, "total_wall_ms": 46200.63591608778, "ttft_ms": 60.15545898117125, "input_usage": {"lookups": 0}, "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": "1.17", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "3d7a51acc5cb98f4e350eed557c0c4a7034a1fdb057a8fc2b38cf2138980b74b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4b9af29f7f94dc18175a7e0f", "attempt_id": "a1870c83e482", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:08.402686Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 506, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 312.9793329862878, "ttft_ms": 25.806166930124164, "decode_ms": 35.36658291704953}, "total_wall_ms": 45865.57787412312, "ttft_ms": 25.806166930124164, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "348152a7613e17b971427cba", "attempt_id": "1c3e42ccc1e4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:08.801027Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 509, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 45490.64429104328, "hops": {"1": 26852.64925006777, "2": 18637.99504097551}, "receiver_prefill_ms": 321.41995802521706, "ttft_ms": 36.68479097541422, "decode_ms": 18.786458065733314}, "total_wall_ms": 45868.29570797272, "ttft_ms": 36.68479097541422, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6c5f6a314e3d0a7f25ae3214", "attempt_id": "495a18423655", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:45.919247Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 496, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 386.58933399710804, "ttft_ms": 134.29083302617073, "decode_ms": 38.08733297046274}, "total_wall_ms": 82762.14316603728, "ttft_ms": 134.29083302617073, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "21f0c9a49afb6bd43b799cf3", "attempt_id": "2f3b232e8929", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:46.504787Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 504, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 326.8197909928858, "ttft_ms": 36.59412497654557, "decode_ms": 40.373832918703556}, "total_wall_ms": 82606.83237505145, "ttft_ms": 36.59412497654557, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "59fbad5bed691520e43526d1", "attempt_id": "8a041e7e83b5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:46.933326Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 500, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 324.8088749824092, "ttft_ms": 37.49904199503362, "decode_ms": 306.11258395947516}, "total_wall_ms": 82871.88412400428, "ttft_ms": 37.49904199503362, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1111cf2153a4a032de46b0ed", "attempt_id": "49a849e31498", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:47.627084Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 499, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 463.0268330220133, "ttft_ms": 36.061667022295296, "decode_ms": 40.128958062268794}, "total_wall_ms": 82742.53670708276, "ttft_ms": 36.061667022295296, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c824ae5912d7a6540e41758a", "attempt_id": "9e4ef9d09e22", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:48.191883Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 507, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 311.4068750292063, "ttft_ms": 23.64479098469019, "decode_ms": 35.81304207909852}, "total_wall_ms": 82574.11995809525, "ttft_ms": 23.64479098469019, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2435a3e9369fe6ba2431f90d", "attempt_id": "b944f3897d88", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:48.587156Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 506, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 312.254499993287, "ttft_ms": 22.45049993507564, "decode_ms": 24.016750045120716}, "total_wall_ms": 82562.03416606877, "ttft_ms": 22.45049993507564, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d6ea660131324c98ceb4043b", "attempt_id": "28833237121b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:48.970312Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 495, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 345.4050419386476, "ttft_ms": 47.59945801924914, "decode_ms": 47.863875050097704}, "total_wall_ms": 82643.99187406525, "ttft_ms": 47.59945801924914, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5cac0daf837a3875013a6e6a", "attempt_id": "c4b4538904d2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:49.436044Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 499, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 433.91445896122605, "ttft_ms": 32.3472919408232, "decode_ms": 43.45112491864711}, "total_wall_ms": 82712.92337402701, "ttft_ms": 32.3472919408232, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7ca752a124c8b0490abe9b3a", "attempt_id": "e58b5638f018", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:49.968925Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 510, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 317.2005829401314, "ttft_ms": 43.33258292172104, "decode_ms": 124.74308395758271}, "total_wall_ms": 82688.40195809025, "ttft_ms": 43.33258292172104, "input_usage": {"lookups": 0}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "81188e569c402eb1338c8e5e", "attempt_id": "e6e0fb2af335", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:50.479266Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 505, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 293.6562920222059, "ttft_ms": 63.15149995498359, "decode_ms": 60.72737497743219}, "total_wall_ms": 82620.97458296921, "ttft_ms": 63.15149995498359, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6b715d99d004c1d0c4846bcf", "attempt_id": "de3d5e5e0b8c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:50.922307Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 517, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 507.6257078908384, "ttft_ms": 36.10850009135902, "decode_ms": 290.89858301449567}, "total_wall_ms": 83037.98608307261, "ttft_ms": 36.10850009135902, "input_usage": {"lookups": 0}, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3d35c1aab1eb7aea4f0625e4", "attempt_id": "795274ef75b6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:51.782587Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 507, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 311.9575829477981, "ttft_ms": 23.05850002449006, "decode_ms": 998.2777499826625}, "total_wall_ms": 83536.52262396645, "ttft_ms": 23.05850002449006, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 21}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "29.7+24.68+20.7=75.08", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "20ee42c223ccaad7a163f5f49cedc190fe17e1e3328a9903f5c6643f332c7be7"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "405da28e04b4ee0c9098cfcd", "attempt_id": "41d71c53c97e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:53.140861Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 504, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 309.89466700702906, "ttft_ms": 36.98625008109957, "decode_ms": 119.62737503927201}, "total_wall_ms": 82669.88533199765, "ttft_ms": 36.98625008109957, "input_usage": {"lookups": 0}, "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": "1.17", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "3d7a51acc5cb98f4e350eed557c0c4a7034a1fdb057a8fc2b38cf2138980b74b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3a40c24ffc6d8ae65b1bd691", "attempt_id": "0d6271af007f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:53.632727Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 506, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 288.4598330128938, "ttft_ms": 46.96037492249161, "decode_ms": 105.23700003977865}, "total_wall_ms": 82643.85362400208, "ttft_ms": 46.96037492249161, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dc8bcc656ee521253f09db8a", "attempt_id": "6f5b2dea7f7f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:26:54.100173Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 509, "message_tokens": 414, "message_utf8_bytes": 1700, "cache_tensor_bytes": null, "serialized_payload_bytes": 1700, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 82202.46491604485, "hops": {"1": 26852.64925006777, "2": 18637.99504097551, "3": 18428.319750004448, "4": 18283.500874997117}, "receiver_prefill_ms": 398.8892090274021, "ttft_ms": 38.48949994426221, "decode_ms": 30.49145801924169}, "total_wall_ms": 82671.2510411162, "ttft_ms": 38.48949994426221, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "3": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}, "4": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1700}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "23fd7a2386f34215f6e64d7360388109746b3d2434f7caecb5aae0a372a3e07d", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ef9e61c6accf5424b5dbf99f", "attempt_id": "d549738d6481", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:21.499694Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 479.4125419575721, "ttft_ms": 24.742458015680313, "decode_ms": 35.95095803029835}, "total_wall_ms": 27441.670916974545, "ttft_ms": 24.742458015680313, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3638196f2eebe7d8d638e07f", "attempt_id": "8d6deab6ba5b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:22.065131Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 300.9690420003608, "ttft_ms": 32.34949999023229, "decode_ms": 351.8158330116421}, "total_wall_ms": 27586.62620803807, "ttft_ms": 32.34949999023229, "input_usage": {"lookups": 0}, "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": "Lot 114-N in Val-d'Or", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.5}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "5fe28a8c11d664e6e2a006121d6f54de2ee2ee7c9089cc8aa533712fb4c4717f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "43e5a783cf67dbc62257ab64", "attempt_id": "2d1bf2e83cc9", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:22.773937Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 314.69000002834946, "ttft_ms": 37.11000003386289, "decode_ms": 56.147375027649105}, "total_wall_ms": 27309.504083939828, "ttft_ms": 37.11000003386289, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a731d83e0d18fca55f1b619b", "attempt_id": "44ef86b79631", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:23.207959Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 412.03491704072803, "ttft_ms": 35.55320901796222, "decode_ms": 32.245000009424984}, "total_wall_ms": 27381.20299996808, "ttft_ms": 35.55320901796222, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4e0bc09dc617f9e2a15c6176", "attempt_id": "44c14de18a6d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:23.711137Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 329.2939580278471, "ttft_ms": 39.205624954774976, "decode_ms": 281.4795000012964}, "total_wall_ms": 27551.746124983765, "ttft_ms": 39.205624954774976, "input_usage": {"lookups": 0}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d88204a0fa42b78305ede44d", "attempt_id": "8e1cdc5dad4e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:24.385243Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 318.96691699512303, "ttft_ms": 59.55079093109816, "decode_ms": 126.36866595130414}, "total_wall_ms": 27406.492750975303, "ttft_ms": 59.55079093109816, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "234b5f6e2864bf84960392fc", "attempt_id": "686a845c15ec", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:24.914553Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 337.8952919738367, "ttft_ms": 31.758583965711296, "decode_ms": 39.00120803155005}, "total_wall_ms": 27310.328959021717, "ttft_ms": 31.758583965711296, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7a131c74ed12ab3ca1e05380", "attempt_id": "0fd6122c1051", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:25.347616Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 296.41387495212257, "ttft_ms": 40.00458400696516, "decode_ms": 24.114083033055067}, "total_wall_ms": 27262.13620894123, "ttft_ms": 40.00458400696516, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5b7ec92ccc9c83357d077457", "attempt_id": "6940823dd9dd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:25.733224Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 289.6338340360671, "ttft_ms": 58.481791988015175, "decode_ms": 585.4088750202209}, "total_wall_ms": 27834.981166990474, "ttft_ms": 58.481791988015175, "input_usage": {"lookups": 0}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2fbc6aa5546c6b15922475e8", "attempt_id": "fa423e58e1af", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:26.690783Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 293.4673751005903, "ttft_ms": 39.19599996879697, "decode_ms": 23.982959100976586}, "total_wall_ms": 27258.035416947678, "ttft_ms": 39.19599996879697, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "62ecc12cd6d69b25ac5e9564", "attempt_id": "e2dac40c86c3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:27.070837Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 460, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 308.1432910403237, "ttft_ms": 22.20362506341189, "decode_ms": 25.189624982886016}, "total_wall_ms": 27257.34825094696, "ttft_ms": 22.20362506341189, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e243d74939efab250c396e0a", "attempt_id": "de29f57df077", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:27.450345Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 295.491874916479, "ttft_ms": 59.3503339914605, "decode_ms": 438.4158750763163}, "total_wall_ms": 27694.783958024345, "ttft_ms": 59.3503339914605, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f86eee769d297be70e4c6249", "attempt_id": "368aef62fb0f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:28.267415Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 290.4355418868363, "ttft_ms": 26.756374980323017, "decode_ms": 167.06683300435543}, "total_wall_ms": 27385.759708937258, "ttft_ms": 26.756374980323017, "input_usage": {"lookups": 0}, "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": "29700", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "21124245547cb94b888c45ccee01be4e6613773a87ea31371ae72634732dfc62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f27a6294c4a6546d3357f23f", "attempt_id": "6cd61bdf8ef5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:28.775600Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 244.06720895785838, "ttft_ms": 65.25533297099173, "decode_ms": 63.05245799012482}, "total_wall_ms": 27273.893709061667, "ttft_ms": 65.25533297099173, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7e6909db7b4f18ac63d3d8b0", "attempt_id": "891afb3975ee", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:29.172118Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26900.86979197804, "hops": {"1": 26900.86979197804}, "receiver_prefill_ms": 419.508708990179, "ttft_ms": 81.08049992006272, "decode_ms": 34.316582954488695}, "total_wall_ms": 27436.581291956827, "ttft_ms": 81.08049992006272, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aaf91b515dd0d1df78f570aa", "attempt_id": "12cffd67e517", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:47.326590Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 285.58408294338733, "ttft_ms": 36.53779102023691, "decode_ms": 32.60170796420425}, "total_wall_ms": 44847.14408300351, "ttft_ms": 36.53779102023691, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b09b60d6343f780388f82b27", "attempt_id": "81253016b563", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:47.705558Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 309.75570797454566, "ttft_ms": 51.82495794724673, "decode_ms": 60.565166058950126}, "total_wall_ms": 44914.811790920794, "ttft_ms": 51.82495794724673, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f98a066152f216c68d622332", "attempt_id": "c0340fd2e8ee", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:48.153978Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 417.3400830477476, "ttft_ms": 28.91583403106779, "decode_ms": 39.791665971279144}, "total_wall_ms": 44978.38966699783, "ttft_ms": 28.91583403106779, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3fc0e267cd5649ed77dc8a22", "attempt_id": "2f37b3fb8b70", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:48.663495Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 301.5398749848828, "ttft_ms": 33.08295796159655, "decode_ms": 31.497417017817497}, "total_wall_ms": 44858.58387500048, "ttft_ms": 33.08295796159655, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "11dc292327a954fa8a531bb3", "attempt_id": "0a4da792cb85", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:49.053499Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 287.252459092997, "ttft_ms": 58.55512502603233, "decode_ms": 549.2697500158101}, "total_wall_ms": 45387.347082956694, "ttft_ms": 58.55512502603233, "input_usage": {"lookups": 0}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2ffc06960f959550d4127df4", "attempt_id": "01672aa2e601", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:49.972159Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 287.2329578967765, "ttft_ms": 25.7120409514755, "decode_ms": 39.4352920120582}, "total_wall_ms": 44844.656624947675, "ttft_ms": 25.7120409514755, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5c843a0751616b00b347189f", "attempt_id": "991582d60e68", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:50.348218Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 287.43566607590765, "ttft_ms": 36.56974993646145, "decode_ms": 24.55916698090732}, "total_wall_ms": 44840.82058293279, "ttft_ms": 36.56974993646145, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d57ad01e2302524b2639e612", "attempt_id": "162231e2fec3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:50.719429Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 257.145292009227, "ttft_ms": 54.98966702725738, "decode_ms": 60.70008303504437}, "total_wall_ms": 44865.24537508376, "ttft_ms": 54.98966702725738, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b760337bdb82f708b9a84994", "attempt_id": "2581c1e6e783", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:51.116379Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 442.2977500362322, "ttft_ms": 33.96295802667737, "decode_ms": 294.58224994596094}, "total_wall_ms": 45263.25762504712, "ttft_ms": 33.96295802667737, "input_usage": {"lookups": 0}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8313cfe5103a5b9d9cfe971e", "attempt_id": "a8717bf4072e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:51.913878Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 294.10675005055964, "ttft_ms": 40.29179096687585, "decode_ms": 25.90620901901275}, "total_wall_ms": 44852.480832953006, "ttft_ms": 40.29179096687585, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "11cec5b75892464418e83ea2", "attempt_id": "634988c1f708", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:52.297924Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 460, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 281.37983300257474, "ttft_ms": 41.28262505400926, "decode_ms": 59.15204097982496}, "total_wall_ms": 44874.33541601058, "ttft_ms": 41.28262505400926, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb882f4bcb86f3096be960fa", "attempt_id": "72674b4a68f1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:52.702787Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 466.4422500645742, "ttft_ms": 35.85962497163564, "decode_ms": 243.76162490807474}, "total_wall_ms": 45238.38145798072, "ttft_ms": 35.85962497163564, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b00578b0c623e31679be5f60", "attempt_id": "7a5388bfb1cf", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:53.472789Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 287.8654169617221, "ttft_ms": 21.38533303514123, "decode_ms": 145.96666593570262}, "total_wall_ms": 44947.51704088412, "ttft_ms": 21.38533303514123, "input_usage": {"lookups": 0}, "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": "29700", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "21124245547cb94b888c45ccee01be4e6613773a87ea31371ae72634732dfc62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ad48bf7aa7b4f7024b3f58ac", "attempt_id": "6b1bfa98634e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:53.950829Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 270.0221670093015, "ttft_ms": 59.920666040852666, "decode_ms": 60.049957945011556}, "total_wall_ms": 44882.23624997772, "ttft_ms": 59.920666040852666, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1ac61c55c9c032dbbc2e7d14", "attempt_id": "9dce6c5a6ead", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:27:54.366480Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 44491.729374974966, "hops": {"1": 26900.86979197804, "2": 17590.859582996927}, "receiver_prefill_ms": 444.46191599126905, "ttft_ms": 35.45299998950213, "decode_ms": 41.061208001337945}, "total_wall_ms": 45013.342540943995, "ttft_ms": 35.45299998950213, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6eb41d91a17037d285d9c9f9", "attempt_id": "ceff13682fcd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:27.924329Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 292.12187498342246, "ttft_ms": 47.251833020709455, "decode_ms": 64.07966697588563}, "total_wall_ms": 77902.71466691047, "ttft_ms": 47.251833020709455, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c49e18753ad0dd182479fd86", "attempt_id": "33aa564d9b6b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:28.355469Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 422.6800419855863, "ttft_ms": 39.70454097725451, "decode_ms": 42.28754201903939}, "total_wall_ms": 78003.80279088859, "ttft_ms": 39.70454097725451, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0274b5f4c1f2c7197057e334", "attempt_id": "098a5694e829", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:28.882740Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 286.93808300886303, "ttft_ms": 22.25804200861603, "decode_ms": 42.616083985194564}, "total_wall_ms": 77850.85641685873, "ttft_ms": 22.25804200861603, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3b687a1523515623df77c00f", "attempt_id": "7ab556179d6e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:29.257374Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 283.9727500686422, "ttft_ms": 28.39495800435543, "decode_ms": 19.224041956476867}, "total_wall_ms": 77830.61712386552, "ttft_ms": 28.39495800435543, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebb4d338b2bdd51c23ea590b", "attempt_id": "bad5733e996b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:29.612057Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 288.05341594852507, "ttft_ms": 61.756041017360985, "decode_ms": 507.0126249920577}, "total_wall_ms": 78356.00624990184, "ttft_ms": 61.756041017360985, "input_usage": {"lookups": 0}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b0575a5353aa783ff0eb0d29", "attempt_id": "db62ac1600bd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:30.494589Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 293.8440409488976, "ttft_ms": 44.674625038169324, "decode_ms": 23.564834031276405}, "total_wall_ms": 77861.25520791393, "ttft_ms": 44.674625038169324, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a48239bcec2c4b45dd4230ff", "attempt_id": "a584d9e3cf91", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:30.879878Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 266.3326660403982, "ttft_ms": 59.90870809182525, "decode_ms": 60.05516601726413}, "total_wall_ms": 77885.49633300863, "ttft_ms": 59.90870809182525, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8679a35844cc61e1be809635", "attempt_id": "ce21912871fd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:31.290889Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 300.24675000458956, "ttft_ms": 139.0367919811979, "decode_ms": 94.38925003632903}, "total_wall_ms": 78033.11549988575, "ttft_ms": 139.0367919811979, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "487204294f938d3937d68596", "attempt_id": "00adc8a41106", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:31.848540Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 304.04854100197554, "ttft_ms": 22.227291949093342, "decode_ms": 308.8234160095453}, "total_wall_ms": 78134.25758294761, "ttft_ms": 22.227291949093342, "input_usage": {"lookups": 0}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "936b1ede1a27713159fd8e3d", "attempt_id": "7b3c90e52945", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:32.508127Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 242.9628330282867, "ttft_ms": 55.35112495999783, "decode_ms": 55.783624993637204}, "total_wall_ms": 77853.26391691342, "ttft_ms": 55.35112495999783, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "04fac3186c3cc676e53b2fc6", "attempt_id": "593c2a351e5b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:32.885901Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 460, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 342.68291702028364, "ttft_ms": 146.23379101976752, "decode_ms": 38.90837507788092}, "total_wall_ms": 78026.98208286893, "ttft_ms": 146.23379101976752, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1cd4c173ba7eb337caf6ada2", "attempt_id": "e40e3d1864dc", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:33.437026Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 304.226583102718, "ttft_ms": 21.991792018525302, "decode_ms": 243.42475004959852}, "total_wall_ms": 78068.80316592287, "ttft_ms": 21.991792018525302, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "289d2b395d9ff5deaeff9024", "attempt_id": "ce87f52cdfe2", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:34.031819Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 251.32062507327646, "ttft_ms": 11.880708974786103, "decode_ms": 249.70858404412866}, "total_wall_ms": 78012.1284169145, "ttft_ms": 11.880708974786103, "input_usage": {"lookups": 0}, "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": "29700", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "21124245547cb94b888c45ccee01be4e6613773a87ea31371ae72634732dfc62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ba9d55b45428299b4a3953ce", "attempt_id": "abe178cbe30f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:34.569531Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 416.36804200243205, "ttft_ms": 31.45729098469019, "decode_ms": 35.18870798870921}, "total_wall_ms": 77982.14033280965, "ttft_ms": 31.45729098469019, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "32c5e824b6abe39568df3e71", "attempt_id": "dda0d1f8b95c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:28:35.076583Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 77498.57670790516, "hops": {"1": 26900.86979197804, "2": 17590.859582996927, "3": 16502.87883298006, "4": 16503.968499950133}, "receiver_prefill_ms": 305.67883304320276, "ttft_ms": 43.26012497767806, "decode_ms": 24.037166964262724}, "total_wall_ms": 77872.18433292583, "ttft_ms": 43.26012497767806, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "3": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "4": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb5b66f34c8abf37f2cc16bc", "attempt_id": "5ea259d82859", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:02.136316Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 479.99470809008926, "ttft_ms": 119.03491604607552, "decode_ms": 44.851165963336825}, "total_wall_ms": 27304.09824894741, "ttft_ms": 119.03491604607552, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3dcba1c07d0430f9432f3e99", "attempt_id": "f1be41916006", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:02.808439Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 384.4288329128176, "ttft_ms": 40.81229108851403, "decode_ms": 28.947707964107394}, "total_wall_ms": 27114.366958034225, "ttft_ms": 40.81229108851403, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6646870a0402a32fa146d2e6", "attempt_id": "00dd22db6360", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:03.287632Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 598, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 380.0157079240307, "ttft_ms": 20.77329100575298, "decode_ms": 227.22233296371996}, "total_wall_ms": 27288.210124941543, "ttft_ms": 20.77329100575298, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e4c27c3d06cee2e74540b818", "attempt_id": "867b8c1c9305", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:03.939632Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 497.99008294939995, "ttft_ms": 47.107750084251165, "decode_ms": 37.29291702620685}, "total_wall_ms": 27242.468249984086, "ttft_ms": 47.107750084251165, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a45b55b4bbd1e910c8da50d2", "attempt_id": "1e49dcebf387", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:04.549873Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 387.05912500154227, "ttft_ms": 36.728749983012676, "decode_ms": 39.056916953995824}, "total_wall_ms": 27123.10658302158, "ttft_ms": 36.728749983012676, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9013035c2107b66cac5a695e", "attempt_id": "6c1c43b777c3", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:05.037941Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 376.46233290433884, "ttft_ms": 17.580208950676024, "decode_ms": 43.82566607091576}, "total_wall_ms": 27098.424957017414, "ttft_ms": 17.580208950676024, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "34298d80be0a9f3774461022", "attempt_id": "809bee92af4a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:05.501987Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 593, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 364.4553329795599, "ttft_ms": 58.57045797165483, "decode_ms": 146.26545901410282}, "total_wall_ms": 27229.374832939357, "ttft_ms": 58.57045797165483, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a222f96dc4307248a8f0edc8", "attempt_id": "0595794812a2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:06.094681Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 597, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 429.69862499739975, "ttft_ms": 24.41954193636775, "decode_ms": 37.79429104179144}, "total_wall_ms": 27151.998540968634, "ttft_ms": 24.41954193636775, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "yes", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "8a798890fe93817163b10b5f7bd2ca4d25d84c52739a645a889c173eee7d9d3d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e7d41af9d742003829219dc", "attempt_id": "ba3d7fedbee8", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:06.609954Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 395.96129197161645, "ttft_ms": 24.270165944471955, "decode_ms": 99.2850000038743}, "total_wall_ms": 27179.615458007902, "ttft_ms": 24.270165944471955, "input_usage": {"lookups": 0}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eac9e5fd68774082b8bd9383", "attempt_id": "ff547cff07e7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:07.153666Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 603, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 364.88054192159325, "ttft_ms": 44.595375075004995, "decode_ms": 70.93633303884417}, "total_wall_ms": 27140.526541043073, "ttft_ms": 44.595375075004995, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "895fa2a16a6536c662e91a8c", "attempt_id": "e206005d9471", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:07.659465Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 615, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 514.9757920298725, "ttft_ms": 29.730957932770252, "decode_ms": 290.63012497499585}, "total_wall_ms": 27495.69262401201, "ttft_ms": 29.730957932770252, "input_usage": {"lookups": 0}, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8a42d9afc9b8daa4dba0fbe3", "attempt_id": "4a6ef89ceab1", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:08.518499Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 605, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 351.4328330056742, "ttft_ms": 38.735541980713606, "decode_ms": 1823.3200830873102}, "total_wall_ms": 28873.68862493895, "ttft_ms": 38.735541980713606, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 38}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "29.7+24.68+20.7+13.3+21.05+1.17=111.20", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "22eda8802dc2949095cea9119990e1f585a51a909516d797763da8b0c1a15061"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a2d9d97fe2c89b3a68a5fbd", "attempt_id": "5e42d2d7598c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:10.756835Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 389.4475830020383, "ttft_ms": 32.76062500663102, "decode_ms": 141.5320000378415}, "total_wall_ms": 27223.746457952075, "ttft_ms": 32.76062500663102, "input_usage": {"lookups": 0}, "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": "1.17", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "3d7a51acc5cb98f4e350eed557c0c4a7034a1fdb057a8fc2b38cf2138980b74b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "409f940ff0c9c2c18c9e9b0a", "attempt_id": "d220628a9387", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:11.343888Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 362.695874995552, "ttft_ms": 53.89670794829726, "decode_ms": 39.21508393250406}, "total_wall_ms": 27115.99858198315, "ttft_ms": 53.89670794829726, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0b7bfe77ed462883a69f6c2a", "attempt_id": "1cb3f320a96a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:11.823949Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 2066, "cache_tensor_bytes": null, "serialized_payload_bytes": 2066, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26659.28429097403, "hops": {"1": 26659.28429097403}, "receiver_prefill_ms": 564.6880419226363, "ttft_ms": 24.28570797201246, "decode_ms": 35.61029105912894}, "total_wall_ms": 27284.787749988027, "ttft_ms": 24.28570797201246, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "fd92d6e8f35e41c0b38eea24a1986a73e8304612f586b8f3be22f50f3e8bf07c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "faabd4a3e215792866241129", "attempt_id": "54a3ed102ee8", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:34.717584Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 572, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 367.15000006370246, "ttft_ms": 38.85274997446686, "decode_ms": 35.877375048585236}, "total_wall_ms": 49342.63991587795, "ttft_ms": 38.85274997446686, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "37cc5bf0ff13d65677ddb16b", "attempt_id": "7107232e387b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:35.183723Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 580, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 384.1712500434369, "ttft_ms": 21.447125007398427, "decode_ms": 27.556374901905656}, "total_wall_ms": 49334.059915854596, "ttft_ms": 21.447125007398427, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4c5addc2de7f6fd094b2f713", "attempt_id": "f4ca93efbcc7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:35.641538Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 576, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 364.2813330516219, "ttft_ms": 37.93541598133743, "decode_ms": 401.20504202786833}, "total_wall_ms": 49704.39174887724, "ttft_ms": 37.93541598133743, "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": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ab98a62731a6d123df87a152", "attempt_id": "7d8826528fc1", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:36.470793Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 575, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 377.1194169530645, "ttft_ms": 21.122499951161444, "decode_ms": 39.024417055770755}, "total_wall_ms": 49338.15008192323, "ttft_ms": 21.122499951161444, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8191d76c959b10ca5cd8b39c", "attempt_id": "ef459a01a639", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:36.931769Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 583, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 348.77829102333635, "ttft_ms": 57.39304097369313, "decode_ms": 546.24337493442}, "total_wall_ms": 49853.241498814896, "ttft_ms": 57.39304097369313, "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": "2772580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "47ef9fbc888d0a8cdba591ce126050123435b1ac90432055aa9598daecdd1504"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b6b6606336e4553350b2cace", "attempt_id": "33d31b791358", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:37.907458Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 582, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 378.64041689317673, "ttft_ms": 40.6536249211058, "decode_ms": 41.6299580829218}, "total_wall_ms": 49361.7749158293, "ttft_ms": 40.6536249211058, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "99126642e1eafa385ddb7013", "attempt_id": "03a8570fb0a7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:38.392851Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 571, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 365.923915989697, "ttft_ms": 18.506749998778105, "decode_ms": 56.993417092598975}, "total_wall_ms": 49342.24829089362, "ttft_ms": 18.506749998778105, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7d2269f37832f5a545cd7281", "attempt_id": "b64a30241220", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:38.858420Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 575, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 349.83387507963926, "ttft_ms": 44.797749957069755, "decode_ms": 134.20145795680583}, "total_wall_ms": 49429.66341599822, "ttft_ms": 44.797749957069755, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a9dc413d3c6f898ebef0cf93", "attempt_id": "c169a87fdb26", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:39.411687Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 586, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 443.08545789681375, "ttft_ms": 30.905249994248152, "decode_ms": 318.18570801988244}, "total_wall_ms": 49693.0407488253, "ttft_ms": 30.905249994248152, "input_usage": {"lookups": 3}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3cdd876ac1278d94e5298089", "attempt_id": "6e67c97d1aa3", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:40.227779Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 581, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 349.4010000722483, "ttft_ms": 41.52745800092816, "decode_ms": 441.1478750407696}, "total_wall_ms": 49732.88145684637, "ttft_ms": 41.52745800092816, "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": "29.32 MW", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "afdffadbe5e60db8a77f2a5b1cbd6f2e15f53b1830505d8a3e7c0af7954bb48b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2a14f5c60a8e08bc922fadc6", "attempt_id": "b90a97bfbd21", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:41.083576Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 593, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 389.849542058073, "ttft_ms": 21.20441710576415, "decode_ms": 284.4295420218259}, "total_wall_ms": 49596.472331904806, "ttft_ms": 21.20441710576415, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2b81f6e4f2c9c96b9bd4cc76", "attempt_id": "cff5316f0d71", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:41.802782Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 583, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 358.22604200802743, "ttft_ms": 35.77637497801334, "decode_ms": 469.0808749292046}, "total_wall_ms": 49763.88987386599, "ttft_ms": 35.77637497801334, "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": "1430223", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "99a5c2c71b999b9018d0009ad96568504eb1c38fcc066245a63abc2387e1c034"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2fa16a1ec18c67dcc6385eaa", "attempt_id": "24e68e72477e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:42.691373Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 580, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 391.1712090484798, "ttft_ms": 39.53554108738899, "decode_ms": 133.46525002270937}, "total_wall_ms": 49464.8854569532, "ttft_ms": 39.53554108738899, "input_usage": {"lookups": 3}, "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": "1170", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "19d0b0d4077454bdec0d553cc96ce822959a5707d28de909cd32395eb0b6f645"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7ff114ec43f0065ca8a23fda", "attempt_id": "90783f3b52b9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:43.279081Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 582, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 342.95104106422514, "ttft_ms": 54.41445799078792, "decode_ms": 60.27541705407202}, "total_wall_ms": 49358.61708188895, "ttft_ms": 54.41445799078792, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ab90df1a894372fb5493057d", "attempt_id": "cfa1db16ae82", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:29:43.760910Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 585, "message_tokens": 490, "message_utf8_bytes": 1539, "cache_tensor_bytes": null, "serialized_payload_bytes": 1539, "side_information_bytes": 0, "fetched_bytes": 1452, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 48900.11099888943, "hops": {"1": 26659.28429097403, "2": 22240.826707915403}, "receiver_prefill_ms": 538.7110830051824, "ttft_ms": 37.74937498383224, "decode_ms": 32.80291601549834}, "total_wall_ms": 49510.22574887611, "ttft_ms": 37.74937498383224, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "642b9d571b19b1cfa27bf501538c4733bd6b5de4fb7ec75282473218ad58ee2f", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "661bbdc88f20196c9d37a1be", "attempt_id": "51f9c2cf8664", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:29.365500Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 578, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 359.72704109735787, "ttft_ms": 36.43991600256413, "decode_ms": 59.34625002555549}, "total_wall_ms": 94323.86904093437, "ttft_ms": 36.43991600256413, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "890d6681ed815a936274ee75", "attempt_id": "9d5c4f344633", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:29.845027Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 586, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 535.7237090356648, "ttft_ms": 40.04583309870213, "decode_ms": 24.80787504464388}, "total_wall_ms": 94468.73574890196, "ttft_ms": 40.04583309870213, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4ca7907cbe326a6280471fbb", "attempt_id": "581f095a39c9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:30.468937Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 582, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 382.4909160612151, "ttft_ms": 40.62191699631512, "decode_ms": 250.68683293648064}, "total_wall_ms": 94541.97770787869, "ttft_ms": 40.62191699631512, "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": "17.2216", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "2ad94847223961d7291d8d336226147c1ee1f17e68d5f69dd1f3bad893d6116a"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "69a6811a3c87515986125d07", "attempt_id": "d7966e84a723", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:31.169558Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 581, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 369.76425000466406, "ttft_ms": 35.74695799034089, "decode_ms": 57.07887501921505}, "total_wall_ms": 94330.76274883933, "ttft_ms": 35.74695799034089, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9fe296b9a2780c210e6cd764", "attempt_id": "cae08b694393", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:31.658731Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 589, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 531.371708959341, "ttft_ms": 37.51320799347013, "decode_ms": 249.03166701551527}, "total_wall_ms": 94686.14049896132, "ttft_ms": 37.51320799347013, "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": "2772580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "47ef9fbc888d0a8cdba591ce126050123435b1ac90432055aa9598daecdd1504"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9c2ae5ece6f5ea44117ab4b2", "attempt_id": "5d2a5e2ebd79", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:32.501563Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 588, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 360.35533307585865, "ttft_ms": 23.725292063318193, "decode_ms": 80.14816592913121}, "total_wall_ms": 94332.53224892542, "ttft_ms": 23.725292063318193, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c98cb5036d2f31dc84b999d0", "attempt_id": "e5ea92a97e66", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:32.993318Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 577, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 347.0649999799207, "ttft_ms": 92.65679097734392, "decode_ms": 142.86579191684723}, "total_wall_ms": 94450.69895801134, "ttft_ms": 92.65679097734392, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "011f14548d7e997890148903", "attempt_id": "55c4e76ff016", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:33.600145Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 581, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 375.31854200642556, "ttft_ms": 32.215457991696894, "decode_ms": 29.870625003241003}, "total_wall_ms": 94305.51291594747, "ttft_ms": 32.215457991696894, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4be2fec548102f6877f060c2", "attempt_id": "5550cabacd3e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:34.060959Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 592, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 375.73049997445196, "ttft_ms": 31.017750035971403, "decode_ms": 352.58204105775803}, "total_wall_ms": 94627.71145789884, "ttft_ms": 31.017750035971403, "input_usage": {"lookups": 9}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "54af4839f60161d74f221fc1", "attempt_id": "c4262b77c925", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:34.846060Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 587, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 520.9650830365717, "ttft_ms": 38.97462505847216, "decode_ms": 304.86062495037913}, "total_wall_ms": 94733.32441598177, "ttft_ms": 38.97462505847216, "input_usage": {"lookups": 9}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7aa96c0cea870628a889d851", "attempt_id": "480ed94aeef5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:35.735376Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 599, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 370.13133300933987, "ttft_ms": 18.050584010779858, "decode_ms": 572.581792017445}, "total_wall_ms": 94828.93854088616, "ttft_ms": 18.050584010779858, "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": "Obyzenlun-18", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "dd4d4b6cbcacda050d50e0bf24b50963f2888f66f17a1ab1b0a9a802e948670b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5ba5b5364b760c8e6506039e", "attempt_id": "eade95905e39", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:36.719032Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 589, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 381.12516701221466, "ttft_ms": 42.56870795506984, "decode_ms": 234.64949999470264}, "total_wall_ms": 94526.58795681782, "ttft_ms": 42.56870795506984, "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": "1439114", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "8537f22786458b3e507df4815fcbd3dae23125b4fece8f1ddfc6900570a1749c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "91d0525dcca2bd7d1b2f39b4", "attempt_id": "ce5a68c4e2d0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:37.402717Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 586, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 352.5180419674143, "ttft_ms": 44.68562500551343, "decode_ms": 199.492167099379}, "total_wall_ms": 94464.86941596959, "ttft_ms": 44.68562500551343, "input_usage": {"lookups": 9}, "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": "1170", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "19d0b0d4077454bdec0d553cc96ce822959a5707d28de909cd32395eb0b6f645"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f05b298d24c9ff963a556c84", "attempt_id": "ba75f196d912", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:38.022554Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 588, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 515.4677910031751, "ttft_ms": 38.932207971811295, "decode_ms": 25.637084036134183}, "total_wall_ms": 94448.23116483167, "ttft_ms": 38.932207971811295, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ee33edef6363a52939c0bfea", "attempt_id": "66662c9da37e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:30:38.628546Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 591, "message_tokens": 496, "message_utf8_bytes": 1551, "cache_tensor_bytes": null, "serialized_payload_bytes": 1551, "side_information_bytes": 0, "fetched_bytes": 4356, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 93867.46187391691, "hops": {"1": 26659.28429097403, "2": 22240.826707915403, "3": 22531.129542039707, "4": 22436.22133298777}, "receiver_prefill_ms": 385.80220798030496, "ttft_ms": 31.46633307915181, "decode_ms": 34.340540994890034}, "total_wall_ms": 94319.92329086643, "ttft_ms": 31.46633307915181, "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": {"ledger": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 23, "retention_rate": 0.36507936507936506, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 2066}, "2": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1539}, "3": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}, "4": {"facts_total": 63, "facts_retained": 30, "retention_rate": 0.47619047619047616, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.3333333333333333, "status": 0.9166666666666666, "place": 0.0, "identifier": 0.5, "quantity": 1.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1551}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "f51608f4d141ff9f08b9709c114c91f613fbd544c41a3467aa7ad1e7d67e2cea", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a23cfbba99038fe33f4a10ac", "attempt_id": "cbd9a4b61c47", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:06.121270Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 286.102749989368, "ttft_ms": 38.09649997856468, "decode_ms": 28.755999985150993}, "total_wall_ms": 27365.908124949783, "ttft_ms": 38.09649997856468, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e4b08bb944b7238d01346b43", "attempt_id": "d27ae81bc944", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:06.498086Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 304.7093750210479, "ttft_ms": 23.723333957605064, "decode_ms": 586.7128750542179}, "total_wall_ms": 27928.253125981428, "ttft_ms": 23.723333957605064, "input_usage": {"lookups": 0}, "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": "Lot 114-N in Val-d'Or", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.5}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "5fe28a8c11d664e6e2a006121d6f54de2ee2ee7c9089cc8aa533712fb4c4717f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "22f899aa7681fa501ec781c0", "attempt_id": "e79a9d30142f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:07.435911Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 338.79241708200425, "ttft_ms": 40.89679196476936, "decode_ms": 30.604249914176762}, "total_wall_ms": 27423.357084044255, "ttft_ms": 40.89679196476936, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8fed8f25c7caa56581201b8a", "attempt_id": "53f443b86797", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:07.869598Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 278.63112499471754, "ttft_ms": 39.68491603154689, "decode_ms": 26.463249931111932}, "total_wall_ms": 27357.872458989732, "ttft_ms": 39.68491603154689, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8d724e97c3dbead297c70a1f", "attempt_id": "3c9b60b268d0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:08.239132Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 267.28945795912296, "ttft_ms": 63.425582950003445, "decode_ms": 592.7645829506218}, "total_wall_ms": 27936.616876046173, "ttft_ms": 63.425582950003445, "input_usage": {"lookups": 0}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4fcada24626ae49411f78923", "attempt_id": "20ab4e662da2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:09.185840Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 289.04662502463907, "ttft_ms": 24.638459086418152, "decode_ms": 38.8843750115484}, "total_wall_ms": 27365.625709062442, "ttft_ms": 24.638459086418152, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "26eb45b6d010006f0ae90740", "attempt_id": "166e28473b91", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:09.563522Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 296.4717079885304, "ttft_ms": 19.033417105674744, "decode_ms": 19.741500029340386}, "total_wall_ms": 27348.185418057255, "ttft_ms": 19.033417105674744, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0a4d81cd77443e71a0154406", "attempt_id": "60d8b402488c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:09.921863Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 442, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 272.7253750199452, "ttft_ms": 64.90666593890637, "decode_ms": 65.8830419415608}, "total_wall_ms": 27416.578334057704, "ttft_ms": 64.90666593890637, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4cdb568ed73bb98aa897782a", "attempt_id": "afffdbc03bef", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:10.350789Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 443.9570000395179, "ttft_ms": 34.95804197154939, "decode_ms": 335.61987499706447}, "total_wall_ms": 27827.524750027806, "ttft_ms": 34.95804197154939, "input_usage": {"lookups": 0}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3644eb8f17b5f8fed34a41b4", "attempt_id": "6050e1ec6123", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:11.190033Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 265.87941707111895, "ttft_ms": 23.919958039186895, "decode_ms": 80.73466701898724}, "total_wall_ms": 27383.668208960444, "ttft_ms": 23.919958039186895, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a80012266749b45c60e41968", "attempt_id": "205cf172c4c7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:11.584631Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 460, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 298.4934999840334, "ttft_ms": 97.30241599027067, "decode_ms": 93.55587500613183}, "total_wall_ms": 27502.397166914307, "ttft_ms": 97.30241599027067, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "37fbb892f49b3fc7b4ac099e", "attempt_id": "425c56761028", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:12.097087Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 350.7881660480052, "ttft_ms": 26.36279200669378, "decode_ms": 253.0728749698028}, "total_wall_ms": 27643.36429303512, "ttft_ms": 26.36279200669378, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e3b71a8aba312c1bf54f2128", "attempt_id": "f2b44fbeb5d6", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:12.752881Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 276.2794999871403, "ttft_ms": 40.19195900764316, "decode_ms": 274.5982090709731}, "total_wall_ms": 27604.104792000726, "ttft_ms": 40.19195900764316, "input_usage": {"lookups": 0}, "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": "29700", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "21124245547cb94b888c45ccee01be4e6613773a87ea31371ae72634732dfc62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ba94a94c61c95852a0da3de5", "attempt_id": "a3339aa06d5a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:13.369201Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 449, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 421.08541692141443, "ttft_ms": 37.152750068344176, "decode_ms": 24.093042011372745}, "total_wall_ms": 27495.34316698555, "ttft_ms": 37.152750068344176, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "727d051873e442231efcf7fa", "attempt_id": "61527c616978", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:13.875484Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 830, "cache_tensor_bytes": null, "serialized_payload_bytes": 830, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27012.44920899626, "hops": {"1": 27012.44920899626}, "receiver_prefill_ms": 302.73908295203, "ttft_ms": 41.969708050601184, "decode_ms": 37.32483310159296}, "total_wall_ms": 27395.04237496294, "ttft_ms": 41.969708050601184, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "458e72ee5799ceb086fadb4f26200c39aa5aa207ba352fb527b3dd31edad8bff", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f1c39ba0e9aebabd32e18e7f", "attempt_id": "63372610c87e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:30.855412Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 444, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 299.5959169929847, "ttft_ms": 32.92345907539129, "decode_ms": 31.377915991470218}, "total_wall_ms": 43949.375166906975, "ttft_ms": 32.92345907539129, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "91393b1bc65e0cd0348eccbe", "attempt_id": "563d40a514e6", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:31.243265Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 325.9386250283569, "ttft_ms": 32.46658307034522, "decode_ms": 23.146333056502044}, "total_wall_ms": 43966.842375928536, "ttft_ms": 32.46658307034522, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "69fa8030f285f60f9f90d76d", "attempt_id": "1fba9996a684", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:31.648833Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 287.93416696134955, "ttft_ms": 38.0287499865517, "decode_ms": 58.401458081789315}, "total_wall_ms": 43969.67000095174, "ttft_ms": 38.0287499865517, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "704ce8e966157ee585d0c3a1", "attempt_id": "14e4ec852943", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:32.059414Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 424.20533404219896, "ttft_ms": 35.4249159572646, "decode_ms": 24.828707915730774}, "total_wall_ms": 44069.84696001746, "ttft_ms": 35.4249159572646, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e9da88b31ff4be35663137aa", "attempt_id": "6645d544408b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:32.567929Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 455, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 305.62687502242625, "ttft_ms": 41.78537498228252, "decode_ms": 277.0654159830883}, "total_wall_ms": 44209.832501015626, "ttft_ms": 41.78537498228252, "input_usage": {"lookups": 1}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0ba0449dcd06dbb43e6dc3fe", "attempt_id": "815b9ef3ef00", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:33.216609Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 454, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 311.6337079554796, "ttft_ms": 60.42733392678201, "decode_ms": 60.33020804170519}, "total_wall_ms": 44017.99229288008, "ttft_ms": 60.42733392678201, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5e1102dd6eb77ff670b33b7c", "attempt_id": "95c9b7e90660", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:33.674560Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 420.35854095593095, "ttft_ms": 32.28954097721726, "decode_ms": 43.692874955013394}, "total_wall_ms": 44081.532042939216, "ttft_ms": 32.28954097721726, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eb0c5b6a4e854481f6142bcf", "attempt_id": "cc635ec96b0f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:34.195870Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 299.1059999912977, "ttft_ms": 38.26995799317956, "decode_ms": 21.06100006494671}, "total_wall_ms": 43943.59625095967, "ttft_ms": 38.26995799317956, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "609435c594fc0cb56c0f67f7", "attempt_id": "ace708171d2c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:34.576921Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 458, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 287.0229589752853, "ttft_ms": 36.811208934523165, "decode_ms": 614.0498330350965}, "total_wall_ms": 44523.28679303173, "ttft_ms": 36.811208934523165, "input_usage": {"lookups": 1}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b2893f1585137b5cac66bc6d", "attempt_id": "ace4ea8996bd", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:35.537442Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 310.5734999990091, "ttft_ms": 41.03962506633252, "decode_ms": 26.40720794443041}, "total_wall_ms": 43963.34887598641, "ttft_ms": 41.03962506633252, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "554344d18f4e13d72ff2fc9b", "attempt_id": "fa2d5d13887d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:35.939178Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 465, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 311.0480000032112, "ttft_ms": 42.003832990303636, "decode_ms": 21.37562504503876}, "total_wall_ms": 43959.880459005944, "ttft_ms": 42.003832990303636, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f35329a278f293515149c84e", "attempt_id": "8c40fe4adfe3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:36.338189Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 455, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 342.7345840027556, "ttft_ms": 35.302374977618456, "decode_ms": 442.25445797201246}, "total_wall_ms": 44405.55687586311, "ttft_ms": 35.302374977618456, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c2337d05947d2fafedc00b6f", "attempt_id": "70c2014c238a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:37.185249Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 302.6779170613736, "ttft_ms": 38.96962502039969, "decode_ms": 137.7267079660669}, "total_wall_ms": 44064.64987597428, "ttft_ms": 38.96962502039969, "input_usage": {"lookups": 1}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c34813d1c91ce9835dfffdc1", "attempt_id": "0bf22353805d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:37.687478Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 454, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 294.1804579459131, "ttft_ms": 60.431750025600195, "decode_ms": 59.49712509755045}, "total_wall_ms": 43999.41000097897, "ttft_ms": 60.431750025600195, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6e216e9077b047736f9173c1", "attempt_id": "27ed4d28de7c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:31:38.125411Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 457, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 523, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 43584.71500093583, "hops": {"1": 27012.44920899626, "2": 16572.26579193957}, "receiver_prefill_ms": 366.5574580663815, "ttft_ms": 109.60029205307364, "decode_ms": 33.41912501491606}, "total_wall_ms": 44094.82066798955, "ttft_ms": 109.60029205307364, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "33e3d0fc1f9bdc3fc744edc3", "attempt_id": "93f8506e052e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:11.485621Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 444, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 281.7440000362694, "ttft_ms": 38.85491704568267, "decode_ms": 34.78145902045071}, "total_wall_ms": 76763.60575086437, "ttft_ms": 38.85491704568267, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8be99f209a815a0ab735b2d8", "attempt_id": "25838b80596c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:11.864146Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 275.4644169472158, "ttft_ms": 37.83979103900492, "decode_ms": 55.86825008504093}, "total_wall_ms": 76777.50862587709, "ttft_ms": 37.83979103900492, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9bb3f64ba753da3cecb9fc7c", "attempt_id": "93c288d15d2d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:12.257719Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 448, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 304.59566693753004, "ttft_ms": 95.2494170051068, "decode_ms": 140.25487506296486}, "total_wall_ms": 76948.47233383916, "ttft_ms": 95.2494170051068, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ada0b05e35640a4131baeb07", "attempt_id": "54ee45cd006e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:12.821684Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 280.1105419639498, "ttft_ms": 40.64858297351748, "decode_ms": 36.04325000196695}, "total_wall_ms": 76764.98391781934, "ttft_ms": 40.64858297351748, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a9f07a1e2d5a28f93979499f", "attempt_id": "9ac3244c9fc5", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:13.202191Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 455, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 326.66308293119073, "ttft_ms": 34.13124999497086, "decode_ms": 255.12654101476073}, "total_wall_ms": 77024.28133483045, "ttft_ms": 34.13124999497086, "input_usage": {"lookups": 3}, "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": "2,772,580", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "dfd7dfe32daffb085bdd75702b46492530076de5b57002f43a92d0340b4e5d62"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d88627857494f5dee1114c15", "attempt_id": "0330fbc09559", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:13.842558Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 454, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 299.076916067861, "ttft_ms": 116.24424997717142, "decode_ms": 137.96666695270687}, "total_wall_ms": 76961.66162588634, "ttft_ms": 116.24424997717142, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "875868d5f9c34cbce62c8c26", "attempt_id": "5a7c2801839c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:14.420294Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 443, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 275.7462919689715, "ttft_ms": 40.052833035588264, "decode_ms": 31.82599996216595}, "total_wall_ms": 76755.95425185747, "ttft_ms": 40.052833035588264, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "069c544f22b0f87352525fa9", "attempt_id": "df030005b582", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:14.792620Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 447, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 275.21574997808784, "ttft_ms": 27.669542003422976, "decode_ms": 39.52537500299513}, "total_wall_ms": 76750.76062686276, "ttft_ms": 27.669542003422976, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb6c63429505279114f90787", "attempt_id": "eaacff9ac18a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:15.160441Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 458, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 289.17616698890924, "ttft_ms": 43.40883390977979, "decode_ms": 563.0449580494314}, "total_wall_ms": 77303.98937582504, "ttft_ms": 43.40883390977979, "input_usage": {"lookups": 3}, "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": "2,308,995", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "58edd3e96aae6822e9516729823cf0882b1294f5ec0c417be0226b6a7a84b028"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b5de749bfc70d7af7f34b66a", "attempt_id": "f4e5e7d74105", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:16.080250Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 453, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 310.00670802313834, "ttft_ms": 41.68270796071738, "decode_ms": 24.100415990687907}, "total_wall_ms": 76784.0367098106, "ttft_ms": 41.68270796071738, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1cf19402dc6fbe5259a52dd6", "attempt_id": "35ecf54bc02d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:16.479182Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 465, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 301.5106249367818, "ttft_ms": 34.8318750038743, "decode_ms": 22.20591704826802}, "total_wall_ms": 76766.90883387346, "ttft_ms": 34.8318750038743, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "67abbafd089e7a2ace446a50", "attempt_id": "b8907a818dc3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:16.862483Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 455, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 324.60125000216067, "ttft_ms": 41.669582948088646, "decode_ms": 461.6827500285581}, "total_wall_ms": 77236.15737585351, "ttft_ms": 41.669582948088646, "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": "5183575", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "cd05cce1120b2084e3bea74abf803fe59e68c8a2ae6af9cb3861d8dceb97d49b"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cbdcde5f20ca91e83f769658", "attempt_id": "e8cbf339ecb9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:17.715025Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 452, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 311.0819160938263, "ttft_ms": 38.794416934251785, "decode_ms": 140.1083329692483}, "total_wall_ms": 76898.25787686277, "ttft_ms": 38.794416934251785, "input_usage": {"lookups": 3}, "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": "29.7", "evidence_artifact": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "7528fc365a095348c9de834851aa3b05d6316fda5575401ce1a3cefe6a79d569"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e08b47de71b5b9c42e31d1b", "attempt_id": "b2250a251241", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:18.229091Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 454, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 289.05733407009393, "ttft_ms": 60.3626670781523, "decode_ms": 58.85283392854035}, "total_wall_ms": 76816.72837689985, "ttft_ms": 60.3626670781523, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a840ed5a36d2da74084d133", "attempt_id": "b438b026d9bb", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:18.662964Z", "task_family": "synthetic", "document_id": "syn_02004_8000", "document_hash": "60ef54e7b129a7bbfac65db771fbf93cecb68ad90c412e5c3ab4051472689429", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02004_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5822, "query_tokens": 457, "message_tokens": 362, "message_utf8_bytes": 846, "cache_tensor_bytes": null, "serialized_payload_bytes": 846, "side_information_bytes": 0, "fetched_bytes": 1569, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76407.67554286867, "hops": {"1": 27012.44920899626, "2": 16572.26579193957, "3": 16437.924249912612, "4": 16385.03629202023}, "receiver_prefill_ms": 416.92154202610254, "ttft_ms": 78.70300009381026, "decode_ms": 31.26437496393919}, "total_wall_ms": 76935.1675848011, "ttft_ms": 78.70300009381026, "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": {"ledger": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "ledger_by_hop": {"1": {"facts_total": 63, "facts_retained": 8, "retention_rate": 0.12698412698412698, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.08333333333333333, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 830}, "2": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "3": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}, "4": {"facts_total": 63, "facts_retained": 13, "retention_rate": 0.20634920634920634, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.1111111111111111, "amount": 0.16666666666666666, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 846}}}, "reference_key_id": "keys/synthetic/syn_02004_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/E4_pilot.log", "payload_hash": "adde1c8ff1e5f75157a5dfec0564662900fc0c278a5331f865c8dea814c0a61c", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "95dec1cdebf281141a538b73", "attempt_id": "e21701ffac8f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:43.275900Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 399.1319169290364, "ttft_ms": 32.830707961693406, "decode_ms": 30.059166019782424}, "total_wall_ms": 24437.95004102867, "ttft_ms": 32.830707961693406, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5f08712d8793b129e54cddd1", "attempt_id": "d9b91ed47979", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:43.757482Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 369.0595000516623, "ttft_ms": 48.10937505681068, "decode_ms": 55.77824998181313}, "total_wall_ms": 24449.272875091992, "ttft_ms": 48.10937505681068, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "da0fa72d17e2b826c134e5ae", "attempt_id": "661c866bddef", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:44.251252Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 534.6387079916894, "ttft_ms": 40.65720795188099, "decode_ms": 234.66383293271065}, "total_wall_ms": 24786.204832955264, "ttft_ms": 40.65720795188099, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "821a6cca1db9a95f213989e1", "attempt_id": "c08ae40f624e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:45.081318Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 377.8246670262888, "ttft_ms": 26.40466601587832, "decode_ms": 591.4460000349209}, "total_wall_ms": 24971.636000089347, "ttft_ms": 26.40466601587832, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "83f0d893e3e2716957f2439c", "attempt_id": "316ae9d48b53", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:46.096692Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 398.9037499995902, "ttft_ms": 43.76133403275162, "decode_ms": 25.44937503989786}, "total_wall_ms": 24444.0878740279, "ttft_ms": 43.76133403275162, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "87684509169a31dc2d2a354c", "attempt_id": "049c0e253167", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:46.584058Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 392.38950004801154, "ttft_ms": 24.620833108201623, "decode_ms": 17.682916950434446}, "total_wall_ms": 24410.664625116624, "ttft_ms": 24.620833108201623, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "27fd67a332e912ed56fb75e3", "attempt_id": "17551625c2d6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:47.037940Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 393.4406249318272, "ttft_ms": 63.52495797909796, "decode_ms": 99.07991695217788}, "total_wall_ms": 24532.22737403121, "ttft_ms": 63.52495797909796, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fe68e43dea96f90d518875b7", "attempt_id": "646367bc3f2b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:47.615378Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 465.71533393580467, "ttft_ms": 35.74629197828472, "decode_ms": 24.94670799933374}, "total_wall_ms": 24502.455292036757, "ttft_ms": 35.74629197828472, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "db16734b0b5ee7e4471ac79a", "attempt_id": "5c12de5a5787", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:48.162397Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 396.64783293846995, "ttft_ms": 40.29704094864428, "decode_ms": 241.06354208197445}, "total_wall_ms": 24654.12687510252, "ttft_ms": 40.29704094864428, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9ba15ed9218b3ba5e8864ead", "attempt_id": "52fdbbb43def", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:48.861181Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 397.58650003932416, "ttft_ms": 104.8796670511365, "decode_ms": 389.2609999747947}, "total_wall_ms": 24867.967291967943, "ttft_ms": 104.8796670511365, "input_usage": {"lookups": 0}, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28617238ec98dc5c07cea608", "attempt_id": "ad8a0dbb53f4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:49.773916Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 613, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 401.00991702638566, "ttft_ms": 41.239499929361045, "decode_ms": 32.54229191225022}, "total_wall_ms": 24450.824124971405, "ttft_ms": 41.239499929361045, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8c01a1863737c71e13b67c26", "attempt_id": "7f9f1119f8d1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:50.269021Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 406.99624991975725, "ttft_ms": 63.28970903996378, "decode_ms": 411.4461250137538}, "total_wall_ms": 24857.862123986706, "ttft_ms": 63.28970903996378, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e6da63d2cfa22ff2bad44a9a", "attempt_id": "6fbe27ff0bb1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:51.171675Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 390.2051249751821, "ttft_ms": 37.250334047712386, "decode_ms": 24.29995802231133}, "total_wall_ms": 24427.733332966454, "ttft_ms": 37.250334047712386, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2026cdd694dcc3bde1525359", "attempt_id": "3697bbff36cf", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:51.643883Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 357.58233291562647, "ttft_ms": 36.054749973118305, "decode_ms": 55.98562490195036}, "total_wall_ms": 24425.600874936208, "ttft_ms": 36.054749973118305, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2448e6dbc516fdea5065e21d", "attempt_id": "959c1dd05e6b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:32:52.113641Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23975.292708026245, "hops": {"1": 23975.292708026245}, "receiver_prefill_ms": 584.68129101675, "ttft_ms": 20.79987491015345, "decode_ms": 41.03741596918553}, "total_wall_ms": 24622.462083003484, "ttft_ms": 20.79987491015345, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3dfc41c104eacea1a66c4fec", "attempt_id": "4abd87675982", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:15.251319Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 589, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 385.90912497602403, "ttft_ms": 18.60316703096032, "decode_ms": 61.1625830642879}, "total_wall_ms": 46911.08033398632, "ttft_ms": 18.60316703096032, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a88af8481a7dde0e20fa78b9", "attempt_id": "4d14ce51bf87", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:15.738037Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 367.08095797803253, "ttft_ms": 34.884791006334126, "decode_ms": 92.2430829377845}, "total_wall_ms": 46939.63487504516, "ttft_ms": 34.884791006334126, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7d373e351ed3c45abd2cd76c", "attempt_id": "d2b77431d587", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:16.253364Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 467.3752919770777, "ttft_ms": 38.50895806681365, "decode_ms": 205.52758406847715}, "total_wall_ms": 47156.77779202815, "ttft_ms": 38.50895806681365, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c149ae305e88701710dbe4a1", "attempt_id": "61e2cdb524da", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:16.987396Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 591, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 364.5579169970006, "ttft_ms": 13.045082916505635, "decode_ms": 580.3962500067428}, "total_wall_ms": 47403.38570799213, "ttft_ms": 13.045082916505635, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "153ac232e06189dfce8b9598", "attempt_id": "02981f9454fa", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:17.964704Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 597, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 387.28866691235453, "ttft_ms": 37.29954094160348, "decode_ms": 25.967500056140125}, "total_wall_ms": 46895.88120800909, "ttft_ms": 37.29954094160348, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9173a838b8aab6f3a4ecc6ee", "attempt_id": "a4fe593b3d0e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:18.434746Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 391.39587501995265, "ttft_ms": 32.6203330187127, "decode_ms": 28.707916964776814}, "total_wall_ms": 46898.02662504371, "ttft_ms": 32.6203330187127, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9fb3793a40a418b67d7a7d6b", "attempt_id": "d175b663a9f5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:18.907177Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 590, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 397.86879206076264, "ttft_ms": 36.218542023561895, "decode_ms": 52.86787508521229}, "total_wall_ms": 46932.400000980124, "ttft_ms": 36.218542023561895, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0bbb8c603065f96d3768c4b8", "attempt_id": "07813d34529c", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:19.415992Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 590, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 539.1037500230595, "ttft_ms": 31.886458047665656, "decode_ms": 35.9331660438329}, "total_wall_ms": 47052.40245803725, "ttft_ms": 31.886458047665656, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ecca71a3699ed0a5d01d1329", "attempt_id": "4fe75a007834", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:20.044245Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 396.4589169481769, "ttft_ms": 22.10766589269042, "decode_ms": 279.78041605092585}, "total_wall_ms": 47143.72725004796, "ttft_ms": 22.10766589269042, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c5b92b2bde62f47c9738fa38", "attempt_id": "c30e72e5d174", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:20.762817Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 389.54908400774, "ttft_ms": 103.91137492842972, "decode_ms": 389.15574993006885}, "total_wall_ms": 47328.346041962504, "ttft_ms": 103.91137492842972, "input_usage": {"lookups": 0}, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1722aa527f253d8ba1c85374", "attempt_id": "68c4b83ada03", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:21.667016Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 395.07204107940197, "ttft_ms": 20.908750011585653, "decode_ms": 42.9699580417946}, "total_wall_ms": 46904.352250974625, "ttft_ms": 20.908750011585653, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2dbdb072be4eee22001be2f6", "attempt_id": "b0da7b9dd912", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:22.145510Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 395.78491589054465, "ttft_ms": 45.51325005013496, "decode_ms": 436.0442089382559}, "total_wall_ms": 47322.862625005655, "ttft_ms": 45.51325005013496, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "52baf4323b8b0723151b214f", "attempt_id": "56849e9b5668", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:23.044932Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 597, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 380.1671249093488, "ttft_ms": 24.646791047416627, "decode_ms": 38.91058301087469}, "total_wall_ms": 46889.15783399716, "ttft_ms": 24.646791047416627, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cc738427f368dde0e225d5be", "attempt_id": "681186195fad", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:23.508631Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 366.5913339937106, "ttft_ms": 40.378083009272814, "decode_ms": 61.220290954224765}, "total_wall_ms": 46913.78487506881, "ttft_ms": 40.378083009272814, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "02a54ef6459752fa80bf571c", "attempt_id": "c2d5051ee543", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:33:23.997886Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 603, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46444.63854201604, "hops": {"1": 23975.292708026245, "2": 22469.34583398979}, "receiver_prefill_ms": 417.0085840160027, "ttft_ms": 144.47683293838054, "decode_ms": 35.70316592231393}, "total_wall_ms": 47042.527792043984, "ttft_ms": 144.47683293838054, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6b0ef1ded1bbe8882ea2296c", "attempt_id": "058c7639eee3", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:10.211929Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 589, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 391.752416966483, "ttft_ms": 39.9212499614805, "decode_ms": 33.22808397933841}, "total_wall_ms": 92504.7833758872, "ttft_ms": 39.9212499614805, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "57ac460b3ab32ea12eef435b", "attempt_id": "149f70d78431", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:10.695821Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 350.15345795545727, "ttft_ms": 37.12670796085149, "decode_ms": 40.557999978773296}, "total_wall_ms": 92467.48112584464, "ttft_ms": 37.12670796085149, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebbaa975f1dbdc58bb50db7d", "attempt_id": "d395e42fffb6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:11.143131Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 561.8182920152321, "ttft_ms": 43.525374960154295, "decode_ms": 242.1851250110194}, "total_wall_ms": 92887.13150087278, "ttft_ms": 43.525374960154295, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c7b1ec9d8df6bef4a4b7d8b6", "attempt_id": "2e91cfeebea4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:12.010703Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 591, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 394.7212079074234, "ttft_ms": 26.218249928206205, "decode_ms": 386.9966660859063}, "total_wall_ms": 92847.69170882646, "ttft_ms": 26.218249928206205, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e9c4c9ee3e6e31a1a929bf9b", "attempt_id": "288a83fe8ff6", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:12.838924Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 597, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 532.118208007887, "ttft_ms": 37.341916002333164, "decode_ms": 27.76116703171283}, "total_wall_ms": 92636.8522499688, "ttft_ms": 37.341916002333164, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c81960cf205d4c792767da29", "attempt_id": "d8d85b0da123", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:13.455106Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 382.56437494419515, "ttft_ms": 24.30849999655038, "decode_ms": 40.065874927677214}, "total_wall_ms": 92486.57033394556, "ttft_ms": 24.30849999655038, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd3dbb354a69d3589f79a21f", "attempt_id": "c92b61da10b2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:13.921119Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 590, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 365.59974995907396, "ttft_ms": 43.500583968125284, "decode_ms": 51.16975004784763}, "total_wall_ms": 92500.01150090247, "ttft_ms": 43.500583968125284, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d576976da2e4f3b4b802c3e3", "attempt_id": "c5cdefa8138f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:14.401162Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 590, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 407.93829201720655, "ttft_ms": 94.4071669364348, "decode_ms": 83.52787501644343}, "total_wall_ms": 92625.70262490772, "ttft_ms": 94.4071669364348, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fdaa52b7e647ce6e3150912d", "attempt_id": "05203a43dd87", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:15.007424Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 377.37399991601706, "ttft_ms": 37.51491603907198, "decode_ms": 236.77941702771932}, "total_wall_ms": 92691.38033385389, "ttft_ms": 37.51491603907198, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f43c25b9b8381f7dee7e8eb7", "attempt_id": "436e9cb7c21e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:15.679239Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 358.60300005879253, "ttft_ms": 44.983166037127376, "decode_ms": 560.8707090141252}, "total_wall_ms": 93004.27912583109, "ttft_ms": 44.983166037127376, "input_usage": {"lookups": 0}, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f0d4b30dc4c0ff1359868943", "attempt_id": "49860c13bd24", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:16.664334Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 389.42608307115734, "ttft_ms": 38.16483402624726, "decode_ms": 30.356207978911698}, "total_wall_ms": 92497.70716694184, "ttft_ms": 38.16483402624726, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cc7e5bb565fcfff81c296385", "attempt_id": "77c599323cfc", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:17.142016Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 365.5868340283632, "ttft_ms": 20.997125073336065, "decode_ms": 547.7081250865012}, "total_wall_ms": 92974.05366785824, "ttft_ms": 20.997125073336065, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bf88668995e680c80900fd8d", "attempt_id": "b77fac34abe7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:18.095511Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 597, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 384.493957972154, "ttft_ms": 33.858333947137, "decode_ms": 29.22062505967915}, "total_wall_ms": 92487.2125008842, "ttft_ms": 33.858333947137, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "35219aa8a0eb30ef03c583a0", "attempt_id": "0a16a3decd50", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:18.561985Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 398.5246659722179, "ttft_ms": 33.41837506741285, "decode_ms": 31.447208020836115}, "total_wall_ms": 92503.17458389327, "ttft_ms": 33.41837506741285, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "853f0a922aa31680e8358f2b", "attempt_id": "60405eb05a87", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:19.045878Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 603, "message_tokens": 507, "message_utf8_bytes": 1103, "cache_tensor_bytes": null, "serialized_payload_bytes": 1103, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92039.00133387651, "hops": {"1": 23975.292708026245, "2": 22469.34583398979, "3": 22571.007124963216, "4": 23023.35566689726}, "receiver_prefill_ms": 416.46079206839204, "ttft_ms": 39.935249951668084, "decode_ms": 52.867084043100476}, "total_wall_ms": 92549.01154187974, "ttft_ms": 39.935249951668084, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1103}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "74f9e0e3d1528334880f1ee3d23066dbab6156429d4aca8c16afafb87c6c31b7", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "414ca290308951aef5fb7a20", "attempt_id": "f0c066764247", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:43.354525Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 441.08037499245256, "ttft_ms": 43.00499998498708, "decode_ms": 24.445457966066897}, "total_wall_ms": 24285.764875006862, "ttft_ms": 43.00499998498708, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "373fbb1155f02b73170b7a1c", "attempt_id": "e32270dcd4ad", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:43.882829Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 445, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 288.56362507212907, "ttft_ms": 39.17458304204047, "decode_ms": 22.922375006601214}, "total_wall_ms": 24127.738749957643, "ttft_ms": 39.17458304204047, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9eeb78e00ea4176ebb620d5c", "attempt_id": "40755bc87f36", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:44.252678Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 289.8686670232564, "ttft_ms": 23.579625063575804, "decode_ms": 19.107208005152643}, "total_wall_ms": 24109.73200004082, "ttft_ms": 23.579625063575804, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1adfdc2163fcd95e0b3330d0", "attempt_id": "4c53b0236a73", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:44.604624Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 440, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 277.3724169237539, "ttft_ms": 44.887083931826055, "decode_ms": 516.1821249639615}, "total_wall_ms": 24615.790667012334, "ttft_ms": 44.887083931826055, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f9a91a7e8744abf6ad3cba8d", "attempt_id": "3d1ae6bf916f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:45.463877Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 446, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 288.4102499810979, "ttft_ms": 21.635167067870498, "decode_ms": 35.75804107822478}, "total_wall_ms": 24123.17208305467, "ttft_ms": 21.635167067870498, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1420f6f5d55d326de7f8fbab", "attempt_id": "3668f9efaeea", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:45.830445Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 293.01200003828853, "ttft_ms": 19.63941694702953, "decode_ms": 55.9974170755595}, "total_wall_ms": 24145.932249957696, "ttft_ms": 19.63941694702953, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "56a144829d5bcd7e388c6268", "attempt_id": "be9d660eaf56", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:46.220953Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 268.7948339153081, "ttft_ms": 46.86445894185454, "decode_ms": 37.40124998148531}, "total_wall_ms": 24130.139459040947, "ttft_ms": 46.86445894185454, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ba1493f20b6df296466aab07", "attempt_id": "dcebd3673071", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:46.596690Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 421.1642500013113, "ttft_ms": 29.480625060386956, "decode_ms": 38.731666980311275}, "total_wall_ms": 24266.55079203192, "ttft_ms": 29.480625060386956, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6b1bd6591e6ffbc65d7b8cc1", "attempt_id": "3b12de56638f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:47.105698Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 456, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 324.7698749182746, "ttft_ms": 39.20350002590567, "decode_ms": 24.082083022221923}, "total_wall_ms": 24165.31362500973, "ttft_ms": 39.20350002590567, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7a015eb1986bf0b3bed1aab0", "attempt_id": "46ac1ef2387a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:47.514334Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 272.25912490393966, "ttft_ms": 60.81091705709696, "decode_ms": 507.1755419485271}, "total_wall_ms": 24617.334375041537, "ttft_ms": 60.81091705709696, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "986c42322a72d3cddbc68e20", "attempt_id": "a1e6738d4013", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:48.373785Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 457, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 328.4824580186978, "ttft_ms": 25.362374959513545, "decode_ms": 39.41845800727606}, "total_wall_ms": 24170.497708953917, "ttft_ms": 25.362374959513545, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6e6e5986cf6a166161bd054f", "attempt_id": "4db281ea8b92", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:48.787284Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 318.2402909733355, "ttft_ms": 36.276415921747684, "decode_ms": 215.82620800472796}, "total_wall_ms": 24347.45208302047, "ttft_ms": 36.276415921747684, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8557819", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "4cab419f6204958711339fe8c75f751c82acb55b10c2d105c5c87c4ea9841c45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "95a1c3ede435d7de244230ba", "attempt_id": "861ef2d20b0e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:49.376590Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 446, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 296.1726669454947, "ttft_ms": 109.30387501139194, "decode_ms": 142.18929095659405}, "total_wall_ms": 24324.826583033428, "ttft_ms": 109.30387501139194, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d9462f8495e1930643489635", "attempt_id": "6bfab87bc4b7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:49.945257Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 296.8470409978181, "ttft_ms": 37.1774589875713, "decode_ms": 27.621250017546117}, "total_wall_ms": 24138.87337502092, "ttft_ms": 37.1774589875713, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3091dc46d82a9c6089317fb6", "attempt_id": "afd728a2844c", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:34:50.326252Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23776.52225003112, "hops": {"1": 23776.52225003112}, "receiver_prefill_ms": 307.5749169802293, "ttft_ms": 40.0104169966653, "decode_ms": 35.203375038690865}, "total_wall_ms": 24160.114708007313, "ttft_ms": 40.0104169966653, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb6cca4f227db37e266ef102", "attempt_id": "209868431a2d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:07.379474Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 302.91679105721414, "ttft_ms": 36.752707906998694, "decode_ms": 36.41695901751518}, "total_wall_ms": 40802.316374960355, "ttft_ms": 36.752707906998694, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "db8762de330e20d7fec5fab0", "attempt_id": "8e4a5ddfb115", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:07.775257Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 457, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 312.19229102134705, "ttft_ms": 71.19558297563344, "decode_ms": 59.84495906159282}, "total_wall_ms": 40869.66545705218, "ttft_ms": 71.19558297563344, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9b24114a8cf9309a447a1408", "attempt_id": "12a90a977624", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:08.240839Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 455, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 444.6994159370661, "ttft_ms": 36.47879196796566, "decode_ms": 30.843500047922134}, "total_wall_ms": 40938.32166597713, "ttft_ms": 36.47879196796566, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "59df2d07461040c4be9ff70f", "attempt_id": "a16e5dc38efb", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:08.774898Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 452, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 303.60558396205306, "ttft_ms": 38.02616603206843, "decode_ms": 308.12225001864135}, "total_wall_ms": 41075.84725006018, "ttft_ms": 38.02616603206843, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "45b856300a52105995f60d4f", "attempt_id": "01769edb0a91", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:09.443490Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 282.4796249624342, "ttft_ms": 61.22587504796684, "decode_ms": 36.654958967119455}, "total_wall_ms": 40806.87854101416, "ttft_ms": 61.22587504796684, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e4bf4e89b73129712bd29ce", "attempt_id": "b65b59e59a71", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:09.844418Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 462, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 441.79720897227526, "ttft_ms": 37.450959091074765, "decode_ms": 24.587125051766634}, "total_wall_ms": 40930.11766602285, "ttft_ms": 37.450959091074765, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dedac6f7516232f67ea3fe28", "attempt_id": "8752bdbaf486", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:10.369504Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 304.34999999124557, "ttft_ms": 42.268041986972094, "decode_ms": 24.458292056806386}, "total_wall_ms": 40797.31574910693, "ttft_ms": 42.268041986972094, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9745bc0679f95ac1c6582e03", "attempt_id": "0fa290e1a4da", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:10.760405Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 301.77812499459833, "ttft_ms": 21.087082917802036, "decode_ms": 35.94049997627735}, "total_wall_ms": 40785.05695797503, "ttft_ms": 21.087082917802036, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "008c0605c52911e9f6383599", "attempt_id": "174acd698139", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:11.139653Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 468, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 286.01316700223833, "ttft_ms": 63.086916925385594, "decode_ms": 48.09291590936482}, "total_wall_ms": 40823.351583094336, "ttft_ms": 63.086916925385594, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6e28f2797ef980102618f667", "attempt_id": "1337e0a3edff", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:11.556508Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 460, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 438.20750003214926, "ttft_ms": 35.26487504132092, "decode_ms": 261.6129170637578}, "total_wall_ms": 41161.39845806174, "ttft_ms": 35.26487504132092, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2eee24c1dc1be6d6ad5f4b53", "attempt_id": "e2e456ee3444", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:12.311798Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 469, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 301.64408299606293, "ttft_ms": 36.31595801562071, "decode_ms": 34.460874972864985}, "total_wall_ms": 40798.657956998795, "ttft_ms": 36.31595801562071, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a8aaa0b8d67bcbf7f68e6f3b", "attempt_id": "402a21fdc5b9", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:12.704576Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 462, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 313.00870806444436, "ttft_ms": 65.46808406710625, "decode_ms": 458.4637500811368}, "total_wall_ms": 41263.121415977366, "ttft_ms": 65.46808406710625, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8528839", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "f8152c4ffd7806f8942183e509a42b4c14f08ec3c476b4573e173dfd4be4f62e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5721b8d656540dbc0c093b63", "attempt_id": "f05d4f4edab1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:13.560322Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 307.4160839896649, "ttft_ms": 38.505957927554846, "decode_ms": 24.45033297408372}, "total_wall_ms": 40796.75299907103, "ttft_ms": 38.505957927554846, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d857f932507f84d28fc43bf4", "attempt_id": "acc9db2210c5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:13.950341Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 463, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 306.20650004129857, "ttft_ms": 24.737666943110526, "decode_ms": 39.59341603331268}, "total_wall_ms": 40796.69595800806, "ttft_ms": 24.737666943110526, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aead03dc91f1e217b78b3391", "attempt_id": "061fdda26b6d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:14.340146Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 464, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40425.6110410206, "hops": {"1": 23776.52225003112, "2": 16649.08879098948}, "receiver_prefill_ms": 319.79612505529076, "ttft_ms": 54.52791694551706, "decode_ms": 49.64741598814726}, "total_wall_ms": 40850.13999894727, "ttft_ms": 54.52791694551706, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8a38c0e29964fcde722b1d59", "attempt_id": "be9c968737df", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:48.023689Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 306.7671660101041, "ttft_ms": 39.478667080402374, "decode_ms": 20.219750003889203}, "total_wall_ms": 74029.67166586313, "ttft_ms": 39.478667080402374, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "892a80e0a106b1978e1e5446", "attempt_id": "05fc54fd6da9", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:48.410597Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 457, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 284.71879195421934, "ttft_ms": 51.9163329154253, "decode_ms": 64.50733297970146}, "total_wall_ms": 74064.64174902067, "ttft_ms": 51.9163329154253, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b5c29f8aecc4152bc51082a9", "attempt_id": "cfe7e1743e06", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:48.833840Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 455, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 466.5626249043271, "ttft_ms": 38.16358302719891, "decode_ms": 33.88599993195385}, "total_wall_ms": 74201.86741603538, "ttft_ms": 38.16358302719891, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c3eea3fec2df1c05463a4f49", "attempt_id": "25e54bac96d3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:49.392022Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 452, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 295.246250112541, "ttft_ms": 31.15412499755621, "decode_ms": 285.8002500142902}, "total_wall_ms": 74275.56649991311, "ttft_ms": 31.15412499755621, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "50f03f90347bbd277b8c4c73", "attempt_id": "c881ef90a20f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:50.024054Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 301.02329095825553, "ttft_ms": 44.246750068850815, "decode_ms": 43.74824999831617}, "total_wall_ms": 74052.73887491785, "ttft_ms": 44.246750068850815, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "19e318838351038f8898c198", "attempt_id": "179ad7b25e92", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:50.435449Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 462, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 459.8047499312088, "ttft_ms": 38.9251250308007, "decode_ms": 31.808833009563386}, "total_wall_ms": 74193.92895698547, "ttft_ms": 38.9251250308007, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "67c7f644753fbd0609f1decc", "attempt_id": "ea768baffba3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:50.989696Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 297.60845797136426, "ttft_ms": 42.08504199050367, "decode_ms": 24.08212493173778}, "total_wall_ms": 74026.99254092295, "ttft_ms": 42.08504199050367, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "093e8ec454d1c98064191cf8", "attempt_id": "06c41e39dedd", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:51.374886Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 306.50470906402916, "ttft_ms": 20.405209041200578, "decode_ms": 19.962207996286452}, "total_wall_ms": 74010.12445788365, "ttft_ms": 20.405209041200578, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "058305e78ac10d8adb452e14", "attempt_id": "11e457a396bb", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:51.741356Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 468, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 302.67070804256946, "ttft_ms": 45.99216696806252, "decode_ms": 61.08562496956438}, "total_wall_ms": 74072.91945791803, "ttft_ms": 45.99216696806252, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5dcb31d6d745115957477929", "attempt_id": "2ac7a7e533c2", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:52.173040Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 460, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 437.56837502587587, "ttft_ms": 35.281874937936664, "decode_ms": 240.4599169967696}, "total_wall_ms": 74376.57466600649, "ttft_ms": 35.281874937936664, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "56650d7c4a7a761446c58534", "attempt_id": "c52309db9b00", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:52.906500Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 469, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 316.85075000859797, "ttft_ms": 24.69616592861712, "decode_ms": 35.050999955274165}, "total_wall_ms": 74039.82449893374, "ttft_ms": 24.69616592861712, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c40abd498401a01991a9deb2", "attempt_id": "b955440d4486", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:53.304514Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 462, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 324.8179170768708, "ttft_ms": 49.09791599493474, "decode_ms": 455.6804579915479}, "total_wall_ms": 74492.82733292785, "ttft_ms": 49.09791599493474, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8528899", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "8e26291f28ee0124796d4c3c7decd23761b2a87e6d80b571eabe1fcde44ebd2e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a16524e80eedf1490c2a774f", "attempt_id": "3f17128a3c6a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:54.156964Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 324.09029197879136, "ttft_ms": 38.13483298290521, "decode_ms": 30.686750076711178}, "total_wall_ms": 74056.14245694596, "ttft_ms": 38.13483298290521, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3b2f5756c213b59c5e697307", "attempt_id": "d271aff1586d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:54.570580Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 463, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 306.90645810682327, "ttft_ms": 22.12029101792723, "decode_ms": 24.942292016930878}, "total_wall_ms": 74017.28274906054, "ttft_ms": 22.12029101792723, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ae4622ed333a92166e5ddb22", "attempt_id": "a2372e16a6f4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:35:54.944366Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 464, "message_tokens": 369, "message_utf8_bytes": 867, "cache_tensor_bytes": null, "serialized_payload_bytes": 867, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 73662.58666594513, "hops": {"1": 23776.52225003112, "2": 16649.08879098948, "3": 16607.98541700933, "4": 16628.990207915194}, "receiver_prefill_ms": 303.6874580429867, "ttft_ms": 52.22408298868686, "decode_ms": 37.61974989902228}, "total_wall_ms": 74056.7209579749, "ttft_ms": 52.22408298868686, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 867}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "233febeeb81a1a69093bbb11f610ddb58dba5bebf634892660db91cfed31d456", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9492872c184befcbb7ab7ff9", "attempt_id": "9177fb038db4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:19.060923Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 553.8503340212628, "ttft_ms": 38.23249996639788, "decode_ms": 35.63608299009502}, "total_wall_ms": 24329.802416963503, "ttft_ms": 38.23249996639788, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9b450017ec2baf1b317ab102", "attempt_id": "162a14c9a972", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:19.711184Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 387.8130830125883, "ttft_ms": 25.380915962159634, "decode_ms": 37.90225007105619}, "total_wall_ms": 24153.222542023286, "ttft_ms": 25.380915962159634, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ada45f2755cdb30128319e19", "attempt_id": "65a268cc2f9f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:20.182177Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 372.84754193387926, "ttft_ms": 37.159667001105845, "decode_ms": 533.6704170331359}, "total_wall_ms": 24645.78408387024, "ttft_ms": 37.159667001105845, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f32b393407d245186566c2b0", "attempt_id": "10a80fb8cc11", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:21.147058Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 389.56066709943116, "ttft_ms": 37.834082962945104, "decode_ms": 302.358791930601}, "total_wall_ms": 24431.882417062297, "ttft_ms": 37.834082962945104, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "856363f05a3ba05ff464fd40", "attempt_id": "28b9b28cc9b2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:21.898238Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 378.97937500383705, "ttft_ms": 41.6633749846369, "decode_ms": 58.9557500788942}, "total_wall_ms": 24181.936083943583, "ttft_ms": 41.6633749846369, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1eaf5d97f795bfffb8686940", "attempt_id": "9216adc8fd00", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:22.397597Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 529.5119170332327, "ttft_ms": 31.535499962046742, "decode_ms": 33.97908294573426}, "total_wall_ms": 24297.051792033017, "ttft_ms": 31.535499962046742, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cb0851a2cf69d68ed8fe7d06", "attempt_id": "3ed97bc39198", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:23.011735Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 390.8851669402793, "ttft_ms": 33.814916969276965, "decode_ms": 44.57545792683959}, "total_wall_ms": 24171.31395789329, "ttft_ms": 33.814916969276965, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6bc8843d729764519067f764", "attempt_id": "1cebd4e13fd4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:23.500125Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 346.18170897010714, "ttft_ms": 35.668207914568484, "decode_ms": 52.28391708806157}, "total_wall_ms": 24136.382666998543, "ttft_ms": 35.668207914568484, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ac340d81d443dcc0a881d05f", "attempt_id": "756b3e4ac241", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:23.954110Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 587.7235420048237, "ttft_ms": 42.85258299205452, "decode_ms": 250.75725000351667}, "total_wall_ms": 24583.343333913945, "ttft_ms": 42.85258299205452, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cec6e8336459ba398b190ea7", "attempt_id": "4e3d95956e8c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:24.856034Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 396.53345895931125, "ttft_ms": 33.50404207594693, "decode_ms": 385.4076670249924}, "total_wall_ms": 24517.755166976713, "ttft_ms": 33.50404207594693, "input_usage": {"lookups": 0}, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4246d71a8ab1699917fce3d3", "attempt_id": "81aa887e39c2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:25.691881Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 613, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 528.5325419390574, "ttft_ms": 38.292584009468555, "decode_ms": 27.806667028926313}, "total_wall_ms": 24296.835166984238, "ttft_ms": 38.292584009468555, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f35245ac574eb5ab05f9ea88", "attempt_id": "e7de3d4bd4a6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:26.306573Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 389.1375840175897, "ttft_ms": 38.748083054088056, "decode_ms": 220.75554099865258}, "total_wall_ms": 24350.765875075012, "ttft_ms": 38.748083054088056, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8238c136d73953bf06f41c61", "attempt_id": "3b8377b3b196", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:26.975218Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 409.8830410512164, "ttft_ms": 97.47345803771168, "decode_ms": 144.87308403477073}, "total_wall_ms": 24354.55875005573, "ttft_ms": 97.47345803771168, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28435d45dd0157006d685ee7", "attempt_id": "ba53d9d7368a", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:27.649359Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 391.00120798684657, "ttft_ms": 38.78258296754211, "decode_ms": 38.29616692382842}, "total_wall_ms": 24170.21779192146, "ttft_ms": 38.78258296754211, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d5061cb5943c334a40b9950f", "attempt_id": "e8c0c047c7a9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:28.143892Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23701.38279197272, "hops": {"1": 23701.38279197272}, "receiver_prefill_ms": 397.89954200387, "ttft_ms": 33.46379101276398, "decode_ms": 19.50245804619044}, "total_wall_ms": 24152.89033402223, "ttft_ms": 33.46379101276398, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfe4fb9830399bca9d3a6768", "attempt_id": "f7a5b1925c21", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:51.530266Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 384.3208330217749, "ttft_ms": 32.32949995435774, "decode_ms": 39.46908295620233}, "total_wall_ms": 47072.00033403933, "ttft_ms": 32.32949995435774, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e96e93dd8f06bc915ad4a2ca", "attempt_id": "c6c1dc35a07f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:52.007206Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 388.50762508809566, "ttft_ms": 23.865832947194576, "decode_ms": 17.071667010895908}, "total_wall_ms": 47045.177082996815, "ttft_ms": 23.865832947194576, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "83e910b3211b6a67f1d76190", "attempt_id": "30aa290eaf67", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:52.456482Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 400.00920894090086, "ttft_ms": 49.3652910226956, "decode_ms": 404.56725005060434}, "total_wall_ms": 47469.70700006932, "ttft_ms": 49.3652910226956, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f4e3785b61981703ecdb15cd", "attempt_id": "c71c1994de06", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:53.331798Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 391.0860000178218, "ttft_ms": 41.01708298549056, "decode_ms": 303.44695795793086}, "total_wall_ms": 47351.45808302332, "ttft_ms": 41.01708298549056, "input_usage": {"lookups": 4}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "999ba823588d73ec37b860ad", "attempt_id": "eba85a1dc8e4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:54.086969Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 375.4773330874741, "ttft_ms": 107.95050009619445, "decode_ms": 144.8826250853017}, "total_wall_ms": 47244.20458311215, "ttft_ms": 107.95050009619445, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3303e4d2c1fc7a22f58f06bb", "attempt_id": "e4034139afc4", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:54.735765Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 385.97249996382743, "ttft_ms": 38.17879094276577, "decode_ms": 26.91816701553762}, "total_wall_ms": 47066.8101250194, "ttft_ms": 38.17879094276577, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b6cfc9fc6e9491272e6eb9e9", "attempt_id": "8cfbad1aed00", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:55.206624Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 389.62270808406174, "ttft_ms": 37.9105000756681, "decode_ms": 24.461708962917328}, "total_wall_ms": 47067.87700008135, "ttft_ms": 37.9105000756681, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ed32b5d3c91bd72913edc3ec", "attempt_id": "27f9d38c1356", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:55.678142Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 364.38295792322606, "ttft_ms": 48.2876249589026, "decode_ms": 55.954999988898635}, "total_wall_ms": 47084.63429205585, "ttft_ms": 48.2876249589026, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28469d0255ef57b274a0d722", "attempt_id": "f8fa46cb5855", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:56.167577Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 530.7640410028398, "ttft_ms": 37.727124989032745, "decode_ms": 242.99108400009573}, "total_wall_ms": 47427.19191603828, "ttft_ms": 37.727124989032745, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e9946921ec35471ee57da94d", "attempt_id": "bf1f233ec703", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:56.998365Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 372.7116669761017, "ttft_ms": 21.881708060391247, "decode_ms": 582.9854999901727}, "total_wall_ms": 47593.442625016905, "ttft_ms": 21.881708060391247, "input_usage": {"lookups": 4}, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b90877e9e65e1cf087bdc6a3", "attempt_id": "200dd2fa1ca3", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:57.995707Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 613, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 403.1182089820504, "ttft_ms": 41.27179202623665, "decode_ms": 32.99441700801253}, "total_wall_ms": 47093.29141606577, "ttft_ms": 41.27179202623665, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "04154e5579717c9947c89815", "attempt_id": "0fe75cb7cbf9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:58.494097Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 392.5395420519635, "ttft_ms": 21.049999981187284, "decode_ms": 351.8786249915138}, "total_wall_ms": 47381.24604197219, "ttft_ms": 21.049999981187284, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "825200f6a749eaaa43b88e77", "attempt_id": "d8d3f3e78b69", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:59.279394Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 518.7579169869423, "ttft_ms": 41.09254095237702, "decode_ms": 19.774917047470808}, "total_wall_ms": 47195.58858405799, "ttft_ms": 41.09254095237702, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e0265c7f03f830740d012ec5", "attempt_id": "ef66f8ff9a16", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:36:59.879599Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 391.05954102706164, "ttft_ms": 36.511249956674874, "decode_ms": 32.81312505714595}, "total_wall_ms": 47076.32225006819, "ttft_ms": 36.511249956674874, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d94b29ae22a4c4cb08bc1b81", "attempt_id": "489779e6da7b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:00.360663Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 861, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46615.0970000308, "hops": {"1": 23701.38279197272, "2": 22913.71420805808}, "receiver_prefill_ms": 357.10641695186496, "ttft_ms": 53.56291704811156, "decode_ms": 60.393083025701344}, "total_wall_ms": 47086.92441601306, "ttft_ms": 53.56291704811156, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "858461d61bff240b47f72bc7", "attempt_id": "f773bd4fe477", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:47.090973Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 594, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 404.53041694127023, "ttft_ms": 34.60033400915563, "decode_ms": 32.094082911498845}, "total_wall_ms": 93323.62704200204, "ttft_ms": 34.60033400915563, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7b8b65982882d7a49d6937ca", "attempt_id": "a57cf96de0b5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:47.582128Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 601, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 359.5435410970822, "ttft_ms": 38.38429192546755, "decode_ms": 56.9622500333935}, "total_wall_ms": 93307.16266599484, "ttft_ms": 38.38429192546755, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ce71bd61eb73583833dd520a", "attempt_id": "93c7576bd9e0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:48.058665Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 599, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 566.3537080399692, "ttft_ms": 39.53283396549523, "decode_ms": 236.61537503357977}, "total_wall_ms": 93694.80558298528, "ttft_ms": 39.53283396549523, "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": "37.0357", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "da856e9315f5534494dcd94197f37dc79881786274216d7efeb9ceb4d546b8f3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "94122ca7e3ed663f7f0753c7", "attempt_id": "0526efc4b69e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:48.921208Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 596, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 383.99629201740026, "ttft_ms": 35.8025839086622, "decode_ms": 346.8078749720007}, "total_wall_ms": 93619.04783290811, "ttft_ms": 35.8025839086622, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6d5a19a7b33b177cd4cfd2d2", "attempt_id": "32b050f16187", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:49.707834Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 540.5042920028791, "ttft_ms": 28.52995798457414, "decode_ms": 43.330457992851734}, "total_wall_ms": 93464.52512498945, "ttft_ms": 28.52995798457414, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "113d8730ab1399cc92278e4e", "attempt_id": "229f6c582775", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:50.339445Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 391.72358298674226, "ttft_ms": 22.13649998884648, "decode_ms": 44.08037499524653}, "total_wall_ms": 93310.39554195013, "ttft_ms": 22.13649998884648, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "09e2b2309114b9e217ae4456", "attempt_id": "d8476408737c", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:50.817049Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 365.1098330738023, "ttft_ms": 59.526124969124794, "decode_ms": 41.21745808515698}, "total_wall_ms": 93318.29929200467, "ttft_ms": 59.526124969124794, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "98e59bb6dbe334dcc8b5d86d", "attempt_id": "9a94f9479e48", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:51.303094Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 595, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 414.2251249868423, "ttft_ms": 149.7137499973178, "decode_ms": 43.56583300977945}, "total_wall_ms": 93459.7328749951, "ttft_ms": 149.7137499973178, "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": "No", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "1ea442a134b2a184bd5d40104401f2a37fbc09ccf3f4bc9da161c6099be3691d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d86a3f4752a1daae9239335e", "attempt_id": "eb7ec1e0b1e1", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:51.930086Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 612, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 392.5440839957446, "ttft_ms": 31.98191709816456, "decode_ms": 249.35083405580372}, "total_wall_ms": 93526.10287501011, "ttft_ms": 31.98191709816456, "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": "1837420", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "de4f2240b3343ced80f9d87227bf2f37a255e07032fcdfc7457f7fbb76e25d0c"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "31ea78ba4dca2146d423113e", "attempt_id": "365c4582c931", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:52.623906Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 604, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 383.70341598056257, "ttft_ms": 60.963124968111515, "decode_ms": 524.6472919825464}, "total_wall_ms": 93821.67737488635, "ttft_ms": 60.963124968111515, "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": "3,833,425", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "bf43ce3c39ad593ef830c4207fbbf3294a04fd70432b5a7bcaaa8dcb77f17ea5"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e901f43a26e1f647aa9e8cd8", "attempt_id": "870b6f6dc208", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:53.613865Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 613, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 383.8153750402853, "ttft_ms": 38.46099996007979, "decode_ms": 23.535249987617135}, "total_wall_ms": 93298.1686249841, "ttft_ms": 38.46099996007979, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "daea2ba81763c3354057dea6", "attempt_id": "dd3580fa7748", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:54.080786Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 606, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 343.8270000042394, "ttft_ms": 51.95404205005616, "decode_ms": 516.9761669822037}, "total_wall_ms": 93765.00258396845, "ttft_ms": 51.95404205005616, "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": "8449767", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "5dcf3a9bfbf82d81de651acc8f6efdb7c850a2a9808e76b884e08c8b75e0f94e"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0da8be93d821c8e8ec5d6644", "attempt_id": "8bbb95b45726", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:55.013315Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 602, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 401.82399994228035, "ttft_ms": 34.124666010029614, "decode_ms": 39.58637500181794}, "total_wall_ms": 93327.85724999849, "ttft_ms": 34.124666010029614, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "032d6b056228b4e704774703", "attempt_id": "a918c4ac1047", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:55.510007Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 607, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 393.2512920582667, "ttft_ms": 23.84104102384299, "decode_ms": 18.098709057085216}, "total_wall_ms": 93287.42854099255, "ttft_ms": 23.84104102384299, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b430013469a258326d9adfad", "attempt_id": "bea184f5ea0d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:37:55.964960Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 608, "message_tokens": 512, "message_utf8_bytes": 1109, "cache_tensor_bytes": null, "serialized_payload_bytes": 1109, "side_information_bytes": 0, "fetched_bytes": 2583, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 92851.5003749635, "hops": {"1": 23701.38279197272, "2": 22913.71420805808, "3": 23092.829999979585, "4": 23143.57337495312}, "receiver_prefill_ms": 377.3999590193853, "ttft_ms": 48.91712497919798, "decode_ms": 55.917875026352704}, "total_wall_ms": 93334.46416689549, "ttft_ms": 48.91712497919798, "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": {"ledger": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "2": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "3": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}, "4": {"facts_total": 65, "facts_retained": 27, "retention_rate": 0.4153846153846154, "retention_by_kind": {"person": 0.5, "date": 0.0, "amount": 0.5625, "status": 0.75, "place": 0.0, "identifier": 0.3333333333333333, "quantity": 0.3333333333333333}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1109}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "52e0b80c5cf2b6b710eeef5fcebeae037996997042fb74c1e3eab0e83598fd43", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8d2e5dce148ddf9269d911ab", "attempt_id": "9a3821e8b917", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:20.255087Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 438, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 410.7520410325378, "ttft_ms": 37.152000004425645, "decode_ms": 21.495166933164}, "total_wall_ms": 24254.074958036654, "ttft_ms": 37.152000004425645, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e99389ccd18eb896190acf87", "attempt_id": "390c74358931", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:20.746670Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 445, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 284.5313750440255, "ttft_ms": 38.46091695595533, "decode_ms": 28.116458910517395}, "total_wall_ms": 24135.781666031107, "ttft_ms": 38.46091695595533, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9348ad50f1c1fa53972f9d97", "attempt_id": "d3bf97aa2945", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:21.117835Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 443, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 285.8792080078274, "ttft_ms": 23.60395807772875, "decode_ms": 19.061624887399375}, "total_wall_ms": 24113.431332050823, "ttft_ms": 23.60395807772875, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfecc15ec1dd2b8757ed1a54", "attempt_id": "4f9282fd998b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:21.467452Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 440, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 283.0027079908177, "ttft_ms": 55.96225010231137, "decode_ms": 543.4428749140352}, "total_wall_ms": 24667.12729097344, "ttft_ms": 55.96225010231137, "input_usage": {"lookups": 0}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "757498eb35bb0f6160e5198d", "attempt_id": "6d47e27347d9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:22.371642Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 446, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 298.48037508782, "ttft_ms": 30.775167047977448, "decode_ms": 41.56729194801301}, "total_wall_ms": 24155.459749046713, "ttft_ms": 30.775167047977448, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1fc6a5229d679173ab3acb42", "attempt_id": "4ad7ca433731", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:22.763157Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 271.9328749226406, "ttft_ms": 66.64237496443093, "decode_ms": 59.29850006941706}, "total_wall_ms": 24182.53483308945, "ttft_ms": 66.64237496443093, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6711ee65d4e641994ae7f03f", "attempt_id": "984cbce0ba32", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:23.180373Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 320.1829169411212, "ttft_ms": 155.68075003102422, "decode_ms": 44.966583023779094}, "total_wall_ms": 24305.492499028333, "ttft_ms": 155.68075003102422, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "52a34b60ba12c78003260ff8", "attempt_id": "af348a7e9d60", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:23.721495Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 439, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 276.58270800020546, "ttft_ms": 24.768374976702034, "decode_ms": 43.20866602938622}, "total_wall_ms": 24129.263999057002, "ttft_ms": 24.768374976702034, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7aa6c55bc8ab7d0718e684f1", "attempt_id": "999e2b2374cd", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:24.085221Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 456, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 311.38608406763524, "ttft_ms": 25.30750003643334, "decode_ms": 35.10324994567782}, "total_wall_ms": 24156.458499957807, "ttft_ms": 25.30750003643334, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "52a359f6e045e4f008b78780", "attempt_id": "0eec30eee2aa", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:24.478491Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 448, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 273.6351250205189, "ttft_ms": 51.88662500586361, "decode_ms": 487.2744169551879}, "total_wall_ms": 24597.554166102782, "ttft_ms": 51.88662500586361, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "08b7dfdb2cb1fbcf9e7fb6e8", "attempt_id": "56dbbc59aa43", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:25.310974Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 457, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 327.1146659972146, "ttft_ms": 38.22999994736165, "decode_ms": 34.56512500997633}, "total_wall_ms": 24184.795457986183, "ttft_ms": 38.22999994736165, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "24f4af87746f0e1cd8d61959", "attempt_id": "c978edb51cc9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:25.731908Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 450, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 303.7567500723526, "ttft_ms": 39.888500003144145, "decode_ms": 301.23029195237905}, "total_wall_ms": 24429.644000018016, "ttft_ms": 39.888500003144145, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8557819", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "4cab419f6204958711339fe8c75f751c82acb55b10c2d105c5c87c4ea9841c45"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "03b320ef7f2ecd529010f30b", "attempt_id": "834913265439", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:26.396208Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 446, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 447.8757919277996, "ttft_ms": 34.70920794643462, "decode_ms": 46.60091700498015}, "total_wall_ms": 24313.750375062227, "ttft_ms": 34.70920794643462, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "57b062897b12bc8358b92b19", "attempt_id": "3b608de59527", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:26.945430Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 451, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 306.96058296598494, "ttft_ms": 21.95525006391108, "decode_ms": 37.886374979279935}, "total_wall_ms": 24151.398874120787, "ttft_ms": 21.95525006391108, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "38c89bd868b950440d67c8f3", "attempt_id": "97f4037ce714", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:27.331493Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 452, "message_tokens": 356, "message_utf8_bytes": 837, "cache_tensor_bytes": null, "serialized_payload_bytes": 837, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 23784.089916036464, "hops": {"1": 23784.089916036464}, "receiver_prefill_ms": 305.3877919446677, "ttft_ms": 29.181707999669015, "decode_ms": 20.724792033433914}, "total_wall_ms": 24139.993750024587, "ttft_ms": 29.181707999669015, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "190d8745b0fccc2f3c2672db692dfc3ed5ba5038f6da54fb0d44f5c7fdfacaef", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6592ea9e0a08a225a74a36b5", "attempt_id": "5ad742c872e9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:44.630370Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 453, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 265.56762494146824, "ttft_ms": 55.80383294727653, "decode_ms": 61.72270805109292}, "total_wall_ms": 41091.281500062905, "ttft_ms": 55.80383294727653, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "91cfcd1da611e44aca3fda27", "attempt_id": "c6c4b8a9dd3b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:45.035906Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 460, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 471.38483298476785, "ttft_ms": 39.389041950926185, "decode_ms": 368.83187503553927}, "total_wall_ms": 41587.78174908366, "ttft_ms": 39.389041950926185, "input_usage": {"lookups": 1}, "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": "Lot 144-F in Val-d'Or", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.5}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "27dc48718952d0803873cb17189adf1b25da1cfa3486be7ddf21f212da22b789"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eb42eb60e7b2d9c37a8037f0", "attempt_id": "08e580b48796", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:45.935696Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 304.9506669631228, "ttft_ms": 21.388708031736314, "decode_ms": 20.056458073668182}, "total_wall_ms": 41054.5240821084, "ttft_ms": 21.388708031736314, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4f9a69a2ea64afc45d3685ab", "attempt_id": "992e9a75005f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:46.301356Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 455, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 354.07587501686066, "ttft_ms": 62.97037494368851, "decode_ms": 460.9323750482872}, "total_wall_ms": 41586.14866610151, "ttft_ms": 62.97037494368851, "input_usage": {"lookups": 1}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dfbacc0cd74e606d46e4f659", "attempt_id": "53b28d95827a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:47.200151Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 461, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 312.94729094952345, "ttft_ms": 40.0783329969272, "decode_ms": 37.2269171057269}, "total_wall_ms": 41098.45466609113, "ttft_ms": 40.0783329969272, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1351b7d39e5cd1d71e550060", "attempt_id": "ce326ee55b5b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:47.610388Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 465, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 296.57962499186397, "ttft_ms": 41.931333020329475, "decode_ms": 44.54775003250688}, "total_wall_ms": 41091.207041055895, "ttft_ms": 41.931333020329475, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8ec1856a00bda03117ce1a18", "attempt_id": "0915090d476b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:48.015259Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 454, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 396.24237501993775, "ttft_ms": 132.850916008465, "decode_ms": 40.413707960397005}, "total_wall_ms": 41277.74129109457, "ttft_ms": 132.850916008465, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "38551bb59081dd295c050f56", "attempt_id": "169957d5bdaf", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:48.606394Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 454, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 308.6740410653874, "ttft_ms": 36.177041940391064, "decode_ms": 25.22225002758205}, "total_wall_ms": 41078.283332055435, "ttft_ms": 36.177041940391064, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfa962d4a6c6617a5718c163", "attempt_id": "82e703beae81", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:48.997353Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 471, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 321.2537500075996, "ttft_ms": 36.801707930862904, "decode_ms": 20.577375078573823}, "total_wall_ms": 41086.87783207279, "ttft_ms": 36.801707930862904, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4810b3548a02eca686b93056", "attempt_id": "26508d71246e", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:49.394987Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 463, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 255.44895802158862, "ttft_ms": 51.06437497306615, "decode_ms": 485.92595802620053}, "total_wall_ms": 41500.6404571468, "ttft_ms": 51.06437497306615, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "65873f468bc534d916119f96", "attempt_id": "d193f429503d", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:50.207522Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 472, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 313.89874999877065, "ttft_ms": 37.98133309464902, "decode_ms": 25.03329108003527}, "total_wall_ms": 41085.17370803747, "ttft_ms": 37.98133309464902, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b46ac8f3936c21d7ab96fbbf", "attempt_id": "68df346f683f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:50.604881Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 465, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 327.0390419056639, "ttft_ms": 27.196165989153087, "decode_ms": 283.2600000547245}, "total_wall_ms": 41345.877583138645, "ttft_ms": 27.196165989153087, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "8530839", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "a48a1f2a6f11982afa0b30d79c299a4b069dbf9f0add8302918e5e558b9c77f2"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "751a55ddc468cd7d467c5671", "attempt_id": "f07139fa2813", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:51.262583Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 461, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 367.6554170669988, "ttft_ms": 114.96779194567353, "decode_ms": 21.25804196111858}, "total_wall_ms": 41212.138041039, "ttft_ms": 114.96779194567353, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b38595604bf5fa206ce0c921", "attempt_id": "ae35d8c99cf8", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:51.789254Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 466, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 299.4812079705298, "ttft_ms": 26.7578330822289, "decode_ms": 41.4998329943046}, "total_wall_ms": 41075.86541597266, "ttft_ms": 26.7578330822289, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "97e3482699dc30b13edd2f78", "attempt_id": "def3d36fd4f4", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:38:52.176710Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 467, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 603, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 40707.560416078195, "hops": {"1": 23784.089916036464, "2": 16923.47050004173}, "receiver_prefill_ms": 304.56537497229874, "ttft_ms": 33.82716595660895, "decode_ms": 29.456458054482937}, "total_wall_ms": 41075.85429109167, "ttft_ms": 33.82716595660895, "input_usage": {"lookups": 1}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bf10e11c2473f93297a3088c", "attempt_id": "bafb297efa82", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:28.761815Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 453, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 299.2594169918448, "ttft_ms": 25.047500035725534, "decode_ms": 36.666958942078054}, "total_wall_ms": 77264.9386250414, "ttft_ms": 25.047500035725534, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "479ac5f424a859bc1934dd19", "attempt_id": "f3761fffd1bd", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:29.142150Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 460, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 290.02429102547467, "ttft_ms": 43.74558397103101, "decode_ms": 611.8961250176653}, "total_wall_ms": 77849.69045803882, "ttft_ms": 43.74558397103101, "input_usage": {"lookups": 3}, "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": "Lot 144-F in Val-d'Or", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.5}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "27dc48718952d0803873cb17189adf1b25da1cfa3486be7ddf21f212da22b789"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ba98dc250809a5ec89b04e19", "attempt_id": "c32368f3430c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:30.110444Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 458, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 329.6320419758558, "ttft_ms": 39.10408401861787, "decode_ms": 20.515082986094058}, "total_wall_ms": 77293.17591607105, "ttft_ms": 39.10408401861787, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4ff0cd4c21f39a91b59c4caf", "attempt_id": "100c495f1bcf", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:30.521161Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 455, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 285.7615420361981, "ttft_ms": 60.12695794925094, "decode_ms": 555.5421250173822}, "total_wall_ms": 77805.50887505524, "ttft_ms": 60.12695794925094, "input_usage": {"lookups": 3}, "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": "Nnamdi Ouellet-Marchand", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "d50065efbde775127fc87c2d04e4d4ea4d1fb8b1f038de510462cef318a17b76"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b4da77097b56989d8305bee3", "attempt_id": "dd03c5f0220f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:31.442472Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 461, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 312.38691695034504, "ttft_ms": 38.55770907830447, "decode_ms": 23.461790988221765}, "total_wall_ms": 77278.49049912766, "ttft_ms": 38.55770907830447, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f6f23c068e5bbb0a994de7fa", "attempt_id": "93896454cee7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:31.837439Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 465, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 309.89979207515717, "ttft_ms": 40.127042098902166, "decode_ms": 20.79012501053512}, "total_wall_ms": 77274.80745804496, "ttft_ms": 40.127042098902166, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "00395b0f3587ef3caa417e9a", "attempt_id": "9cfcb02774a7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:32.228543Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 454, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 321.5562499826774, "ttft_ms": 60.00966590363532, "decode_ms": 44.2490839632228}, "total_wall_ms": 77329.82220710255, "ttft_ms": 60.00966590363532, "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": "no", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_2000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2d2c05a23a9a5180ba4ad4af", "attempt_id": "8a1957d7044f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:32.676750Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 454, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 452.5126660009846, "ttft_ms": 35.22349998820573, "decode_ms": 27.20854093786329}, "total_wall_ms": 77418.86658302974, "ttft_ms": 35.22349998820573, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9bad38f49df4ad830bcb1501", "attempt_id": "85b36e95f17c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:33.211515Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 471, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 296.8220409238711, "ttft_ms": 37.31437504757196, "decode_ms": 31.771917012520134}, "total_wall_ms": 77269.79070808738, "ttft_ms": 37.31437504757196, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ca52218613e4fc5b7fc499bc", "attempt_id": "532b747baab3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:33.596099Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 463, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 301.82787496596575, "ttft_ms": 20.29408293310553, "decode_ms": 339.20983399730176}, "total_wall_ms": 77565.35195803735, "ttft_ms": 20.29408293310553, "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": "4216925", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "027953cde36b43114e29559555e0c2d2beb78199e351c7188aa92b44d9823854"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fad5e49f5fa076ca20dcb9d6", "attempt_id": "bea456bbefda", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:34.276628Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 472, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 442.995083052665, "ttft_ms": 35.26204207446426, "decode_ms": 29.538333998061717}, "total_wall_ms": 77411.99874912854, "ttft_ms": 35.26204207446426, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b88b053e84b3146c912e9718", "attempt_id": "25f59857ddc2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:34.804038Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 465, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 322.80958304181695, "ttft_ms": 34.106082981452346, "decode_ms": 235.91241601388901}, "total_wall_ms": 77496.79045716766, "ttft_ms": 34.106082981452346, "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": "8529849", "evidence_artifact": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "3681cd52b58373d704c6a432e1a0e05bb362f26f2ad17b3c124e5fa7341b9e5f"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7e48d89397bf7a990ee67852", "attempt_id": "278ba1c07388", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:35.416601Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 461, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 296.9112499849871, "ttft_ms": 40.66120891366154, "decode_ms": 35.72479204740375}, "total_wall_ms": 77277.5539989816, "ttft_ms": 40.66120891366154, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2daebff483cc4ff502213448", "attempt_id": "50dbe7f4ec65", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:35.811642Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 466, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 465.5850420240313, "ttft_ms": 39.490666007623076, "decode_ms": 42.36558300908655}, "total_wall_ms": 77451.49233308621, "ttft_ms": 39.490666007623076, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aaeef541bcdbc4ead6e22f49", "attempt_id": "c32cd3fa6fb1", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:36.379498Z", "task_family": "synthetic", "document_id": "syn_02006_2000", "document_hash": "fe6d5d7c5ffa3ddfdaf46c3cca5a6b9f51c19c0e9987b820daf4e669ad1b48c1", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02006_2000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 2032, "query_tokens": 467, "message_tokens": 372, "message_utf8_bytes": 870, "cache_tensor_bytes": null, "serialized_payload_bytes": 870, "side_information_bytes": 0, "fetched_bytes": 1809, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 76903.38954108302, "hops": {"1": 23784.089916036464, "2": 16923.47050004173, "3": 16950.1928340178, "4": 19245.636290987022}, "receiver_prefill_ms": 314.1344579635188, "ttft_ms": 37.339917034842074, "decode_ms": 30.190291930921376}, "total_wall_ms": 77285.77562409919, "ttft_ms": 37.339917034842074, "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": {"ledger": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "ledger_by_hop": {"1": {"facts_total": 65, "facts_retained": 9, "retention_rate": 0.13846153846153847, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.25, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 837}, "2": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "3": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}, "4": {"facts_total": 65, "facts_retained": 12, "retention_rate": 0.18461538461538463, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.125, "status": 0.5, "place": 0.16666666666666666, "identifier": 0.08333333333333333, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 870}}}, "reference_key_id": "keys/synthetic/syn_02006_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/E4_pilot.log", "payload_hash": "c8a623e44e2b9618463a709caafaabe2c9ec0e15836628af4a2920cf13b0c047", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fefc039576d099112c9c6828", "attempt_id": "50bdd2a2797e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:54.037905Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 380, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 265.963957994245, "ttft_ms": 22.45091600343585, "decode_ms": 20.35129105206579}, "total_wall_ms": 17487.825707998127, "ttft_ms": 22.45091600343585, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "54262efbc7173d16e8970de7", "attempt_id": "6532157f7a08", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:54.371979Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 306.07745808083564, "ttft_ms": 50.80133303999901, "decode_ms": 80.30970802064985}, "total_wall_ms": 17616.19091697503, "ttft_ms": 50.80133303999901, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2dbce45ab2de623969353510", "attempt_id": "7f37697b8d37", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:54.832646Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 386, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 377.6774170110002, "ttft_ms": 32.39649999886751, "decode_ms": 36.68875002767891}, "total_wall_ms": 17625.866999966092, "ttft_ms": 32.39649999886751, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "17763fdb2718ba7daf2d5479", "attempt_id": "21390f3222fe", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:55.303638Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 270.3487080289051, "ttft_ms": 23.025291971862316, "decode_ms": 40.49695795401931}, "total_wall_ms": 17513.02370801568, "ttft_ms": 23.025291971862316, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b72b06877c710f0dcf267ae6", "attempt_id": "0c3e567af8f1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:55.665955Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 268.08904099743813, "ttft_ms": 16.159916995093226, "decode_ms": 11.730958940461278}, "total_wall_ms": 17474.908042000607, "ttft_ms": 16.159916995093226, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "712368dafe5789d62f9acd76", "attempt_id": "a4f68d0406c4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:55.986684Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 389, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 314.39295795280486, "ttft_ms": 51.825584028847516, "decode_ms": 36.18012508377433}, "total_wall_ms": 17581.507041002624, "ttft_ms": 51.825584028847516, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "58b78eafc64ce2e12ac34378", "attempt_id": "a365dac0ef66", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:56.416246Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 409.85029202420264, "ttft_ms": 23.99612497538328, "decode_ms": 38.854500045999885}, "total_wall_ms": 17651.62637399044, "ttft_ms": 23.99612497538328, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aa300cea135fc4c7755a1b28", "attempt_id": "d583f6b6425a", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:56.913337Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 282.68337494228035, "ttft_ms": 39.317250018939376, "decode_ms": 23.457709001377225}, "total_wall_ms": 17524.31195799727, "ttft_ms": 39.317250018939376, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "325a26fa23cf4a5f1f107b45", "attempt_id": "193e5ce01890", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:57.282468Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 397, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 266.62400003988296, "ttft_ms": 55.09645794518292, "decode_ms": 43.8428339548409}, "total_wall_ms": 17544.91620801855, "ttft_ms": 55.09645794518292, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b26f8f8da24dbea7d5b3ce74", "attempt_id": "766fe430296b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:57.671736Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 389, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 278.20358297321945, "ttft_ms": 107.43116599041969, "decode_ms": 149.07358295749873}, "total_wall_ms": 17713.828457985073, "ttft_ms": 107.43116599041969, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fb530aa66bea30de1062902e", "attempt_id": "46c75c45bd14", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:58.230126Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 400, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 273.08058296330273, "ttft_ms": 37.71958302240819, "decode_ms": 134.3756669666618}, "total_wall_ms": 17624.17904101312, "ttft_ms": 37.71958302240819, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b447e65729f1e52491bc2008", "attempt_id": "74ccabc88e3e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:58.698951Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 392, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 290.0840830989182, "ttft_ms": 36.96045896504074, "decode_ms": 30.42974998243153}, "total_wall_ms": 17536.552541074343, "ttft_ms": 36.96045896504074, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9e00b51e8193e5b34f0c12ae", "attempt_id": "df478cca2e52", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:59.081904Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 293.1739999912679, "ttft_ms": 66.59745902288705, "decode_ms": 40.68291699513793}, "total_wall_ms": 17579.94329102803, "ttft_ms": 66.59745902288705, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "499c62061d34ba7f58465eb2", "attempt_id": "66f25dd2b49d", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:39:59.507623Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 392, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 417.93420806061476, "ttft_ms": 35.87537503335625, "decode_ms": 27.720791986212134}, "total_wall_ms": 17660.555458045565, "ttft_ms": 35.87537503335625, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1bb8bdd542dd9b9d75830857", "attempt_id": "cc9332383d21", "run_group_id": "E4_pilot:A4|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:00.014739Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 394, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17178.316958015785, "hops": {"1": 17178.316958015785}, "receiver_prefill_ms": 280.0427919719368, "ttft_ms": 40.21529108285904, "decode_ms": 24.611374945379794}, "total_wall_ms": 17523.9352080971, "ttft_ms": 40.21529108285904, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a9adec97e0efd4c7f4e6b14", "attempt_id": "dd34092e53e5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:10.711233Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 311, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 239.4002500222996, "ttft_ms": 30.2219579461962, "decode_ms": 30.64099990297109}, "total_wall_ms": 27803.535083075985, "ttft_ms": 30.2219579461962, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fc7c6bf868940347a9b8d1d2", "attempt_id": "41c21039a5fc", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:11.034922Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 210.60341701377183, "ttft_ms": 59.7166660008952, "decode_ms": 60.126875061541796}, "total_wall_ms": 27833.702916977927, "ttft_ms": 59.7166660008952, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bb957451429fa791b7e6f97a", "attempt_id": "697fb29bbed4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:11.389845Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 317, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 285.527458996512, "ttft_ms": 123.35220898967236, "decode_ms": 36.7617909796536}, "total_wall_ms": 27949.18995804619, "ttft_ms": 123.35220898967236, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "21c842bfdc0e4f10a65c478d", "attempt_id": "9b6d38b7d680", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:11.862764Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 244.9809170793742, "ttft_ms": 23.873334052041173, "decode_ms": 37.87641599774361}, "total_wall_ms": 27809.990792069584, "ttft_ms": 23.873334052041173, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d63020edff6b41e40afc71a4", "attempt_id": "370bddf30f85", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:12.195580Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 226.04120790492743, "ttft_ms": 41.73670802265406, "decode_ms": 34.77295802440494}, "total_wall_ms": 27805.874624988064, "ttft_ms": 41.73670802265406, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3b42b4daf6981d6f460ccd08", "attempt_id": "53a3bb48ff14", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:12.521222Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 320, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 237.78458405286074, "ttft_ms": 48.26587496791035, "decode_ms": 50.81554199568927}, "total_wall_ms": 27840.35895904526, "ttft_ms": 48.26587496791035, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "938dd094e3c186d795ba9b6e", "attempt_id": "7047bc686492", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:12.884926Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 385.5123749235645, "ttft_ms": 46.507708029821515, "decode_ms": 38.12704200390726}, "total_wall_ms": 27973.40245905798, "ttft_ms": 46.507708029821515, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dafbd707ce507aecd4c1e78d", "attempt_id": "81f2715f37a4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:13.380558Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 266.2227919790894, "ttft_ms": 21.643916028551757, "decode_ms": 43.35333302151412}, "total_wall_ms": 27834.480250137858, "ttft_ms": 21.643916028551757, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2e5b38f3d5ec4cc8510a47fe", "attempt_id": "08185ee48760", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:13.735113Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 328, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 251.11958291381598, "ttft_ms": 28.89845904428512, "decode_ms": 30.51320801023394}, "total_wall_ms": 27813.82279098034, "ttft_ms": 28.89845904428512, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "596c652eb7a06532641ff2e3", "attempt_id": "d9f74c0b3856", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:14.071315Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 320, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 247.36479204148054, "ttft_ms": 57.099209050647914, "decode_ms": 62.75429192464799}, "total_wall_ms": 27870.697875041515, "ttft_ms": 57.099209050647914, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fdb028348239f2d89fb9ffd4", "attempt_id": "30957374cd54", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:14.465766Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 331, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 419.92058302275836, "ttft_ms": 29.93462502490729, "decode_ms": 145.29487502295524}, "total_wall_ms": 28098.405875032768, "ttft_ms": 29.93462502490729, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d198f461875ca88d160a51aa", "attempt_id": "34ced9cff60e", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:15.086573Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 323, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 263.6254579992965, "ttft_ms": 38.37049996946007, "decode_ms": 34.44470907561481}, "total_wall_ms": 27839.686084073037, "ttft_ms": 38.37049996946007, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "de7f8543f027f420ba0c1977", "attempt_id": "ed8a5f8de882", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:15.448238Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 209.88412504084408, "ttft_ms": 37.69262495916337, "decode_ms": 50.69079203531146}, "total_wall_ms": 27801.493000006303, "ttft_ms": 37.69262495916337, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7aded81278cd281630393b73", "attempt_id": "15ab91f4bab5", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:15.769603Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 323, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 240.32858293503523, "ttft_ms": 49.316041986458, "decode_ms": 123.6977499211207}, "total_wall_ms": 27916.743167093955, "ttft_ms": 49.316041986458, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "70e4d2922bae2e9a5a50aabc", "attempt_id": "7337b3547248", "run_group_id": "E4_pilot:A4|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:16.208244Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 325, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27502.637000055984, "hops": {"1": 17178.316958015785, "2": 10324.320042040199}, "receiver_prefill_ms": 315.59479201678187, "ttft_ms": 23.973874980583787, "decode_ms": 39.56370893865824}, "total_wall_ms": 27882.53995811101, "ttft_ms": 23.973874980583787, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "53fe6549f6b23220815748fa", "attempt_id": "2532cc678fe7", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:35.790259Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 298, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 310.1268749451265, "ttft_ms": 37.01812494546175, "decode_ms": 30.81520809791982}, "total_wall_ms": 47055.48504216131, "ttft_ms": 37.01812494546175, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9a3788da6df316a9ff5ae85c", "attempt_id": "7c4c28b0a243", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:36.191402Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 232.4832499725744, "ttft_ms": 30.274250078946352, "decode_ms": 41.724166949279606}, "total_wall_ms": 46981.974292197265, "ttft_ms": 30.274250078946352, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "50ec61e7a5fccc3367b0b4de", "attempt_id": "6a42d9af04ae", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:36.518489Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 304, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 227.8750839177519, "ttft_ms": 19.98649991583079, "decode_ms": 79.13308392744511}, "total_wall_ms": 47004.442209145054, "ttft_ms": 19.98649991583079, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e7e620923b0231517b4f6c87", "attempt_id": "2b8aac3a2389", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:36.869021Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 250.03270793240517, "ttft_ms": 111.93320795428008, "decode_ms": 143.58404208905995}, "total_wall_ms": 47182.96050012577, "ttft_ms": 111.93320795428008, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7bfc1622a20c0c7a8350b653", "attempt_id": "375eef10a5dd", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:37.398791Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 235.49024993553758, "ttft_ms": 40.02687509637326, "decode_ms": 25.15945804771036}, "total_wall_ms": 46978.33354212344, "ttft_ms": 40.02687509637326, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "da783a236586368e7a0505b3", "attempt_id": "b8f43b334112", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:37.722945Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 307, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 238.54108294472098, "ttft_ms": 26.472915895283222, "decode_ms": 41.250749956816435}, "total_wall_ms": 46983.638709178194, "ttft_ms": 26.472915895283222, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7e5042a5d451c74eac4512a9", "attempt_id": "8ead52e8396b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:38.052330Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 216.76025004126132, "ttft_ms": 43.468625051900744, "decode_ms": 59.91433304734528}, "total_wall_ms": 46997.61945917271, "ttft_ms": 43.468625051900744, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "17056a64b6e97b7ab65183ab", "attempt_id": "1fd1839b6c0f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:38.395870Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 235.2721249917522, "ttft_ms": 108.07283397298306, "decode_ms": 123.75562498345971}, "total_wall_ms": 47144.553209189326, "ttft_ms": 108.07283397298306, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ad6a363a46eac8a64a4c8c0a", "attempt_id": "47e20652629b", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:38.887266Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 315, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 226.44483298063278, "ttft_ms": 34.12908304017037, "decode_ms": 36.66700003668666}, "total_wall_ms": 46974.728334113024, "ttft_ms": 34.12908304017037, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "255cd183cde3f41c795f1ab5", "attempt_id": "b9cbf119ade1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:39.207326Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 307, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 213.5139579186216, "ttft_ms": 38.16837503109127, "decode_ms": 33.08387496508658}, "total_wall_ms": 46962.17941713985, "ttft_ms": 38.16837503109127, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd7cfea31893d1bdca613f5c", "attempt_id": "18d3de60794f", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:39.515022Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 318, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 209.19445902109146, "ttft_ms": 21.283165900968015, "decode_ms": 180.8318750699982}, "total_wall_ms": 47089.03587516397, "ttft_ms": 21.283165900968015, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "be1fa7edc03f0bf88d525ffb", "attempt_id": "18a5ec3896b1", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:39.951056Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 310, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 414.33595796115696, "ttft_ms": 32.92845899704844, "decode_ms": 37.110332981683314}, "total_wall_ms": 47161.99125116691, "ttft_ms": 32.92845899704844, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d58a0a50d0e3bd52335fe9d2", "attempt_id": "cf7ed95cacbc", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:40.461217Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 251.21354102157056, "ttft_ms": 24.75191606208682, "decode_ms": 40.692999958992004}, "total_wall_ms": 46994.12883317564, "ttft_ms": 24.75191606208682, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9d7e6b258e07736d4de03273", "attempt_id": "dfddec491ac4", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:40.801757Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 310, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 227.53154207020998, "ttft_ms": 41.119541972875595, "decode_ms": 34.529499942436814}, "total_wall_ms": 46980.614376137964, "ttft_ms": 41.119541972875595, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5bfb24eadc083d8863ad1e1f", "attempt_id": "811a0e5641c2", "run_group_id": "E4_pilot:A4|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:40:41.127424Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46676.90504214261, "hops": {"1": 17178.316958015785, "2": 10324.320042040199, "3": 9517.005792004056, "4": 9657.262250082567}, "receiver_prefill_ms": 235.3209590073675, "ttft_ms": 67.51745799556375, "decode_ms": 60.78075000550598}, "total_wall_ms": 47041.15283314604, "ttft_ms": 67.51745799556375, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebcf3d2cae5c6d3d1119c51f", "attempt_id": "de79bb503a0e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:08.508907Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 414, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 278.5970829427242, "ttft_ms": 20.68829198833555, "decode_ms": 20.186583045870066}, "total_wall_ms": 27290.566334035248, "ttft_ms": 20.68829198833555, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d0af3b59e54f73c794ffeada", "attempt_id": "c2d85315b170", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:08.853690Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 327.1927499445155, "ttft_ms": 55.8269580360502, "decode_ms": 553.9902909658849}, "total_wall_ms": 27908.14841701649, "ttft_ms": 55.8269580360502, "input_usage": {"lookups": 0}, "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": "Lot 697-W in La Tuque", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.8000000000000002}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "ad0a49b81076edddfe98ccb0a2aac3fc27477fd4125ee7370c410ad68686bd8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "79af2933c67824b7b7ecd13e", "attempt_id": "93401831be30", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:09.817540Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 420, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 301.61195900291204, "ttft_ms": 37.765958928503096, "decode_ms": 176.2929999968037}, "total_wall_ms": 27487.003916990943, "ttft_ms": 37.765958928503096, "input_usage": {"lookups": 0}, "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": "2.7230", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "7e908e02eb04c1aa8a80abf6fdd1511a7fa4f87b430c0f0452f47b4063b159f0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "eced06e9c141bc9b51df741a", "attempt_id": "0ff51d87aced", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:10.358411Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 329.66608402784914, "ttft_ms": 104.15975004434586, "decode_ms": 294.05495803803205}, "total_wall_ms": 27699.0638760617, "ttft_ms": 104.15975004434586, "input_usage": {"lookups": 0}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "20c6db0cdcc7b59ff6b5e90a", "attempt_id": "278da633bfb4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:11.111295Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 295.19708396401256, "ttft_ms": 25.636708945967257, "decode_ms": 36.924750078469515}, "total_wall_ms": 27328.739001066424, "ttft_ms": 25.636708945967257, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "879cd945ac1dc492ba985d14", "attempt_id": "e77735c42491", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:11.492360Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 423, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 291.1542080109939, "ttft_ms": 23.492750013247132, "decode_ms": 16.771333990618587}, "total_wall_ms": 27302.479750011116, "ttft_ms": 23.492750013247132, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "57a8970d28b3eceef635aa4e", "attempt_id": "758d4ed90d8a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:11.848697Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 281.5188330132514, "ttft_ms": 53.72212501242757, "decode_ms": 63.57954198028892}, "total_wall_ms": 27370.097376056947, "ttft_ms": 53.72212501242757, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "68939d1a1a392c31b4abb798", "attempt_id": "ff3dd1c1c5ac", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:12.273714Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 422.0522920368239, "ttft_ms": 34.55466695595533, "decode_ms": 38.558749947696924}, "total_wall_ms": 27466.364709078334, "ttft_ms": 34.55466695595533, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3575f31bbe5d11ecf9811907", "attempt_id": "ce377127c021", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:12.793807Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 431, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 295.82804103847593, "ttft_ms": 32.89525001309812, "decode_ms": 31.475541996769607}, "total_wall_ms": 27331.38779201545, "ttft_ms": 32.89525001309812, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "be3db4f20aaca7df08952cb4", "attempt_id": "714d735c591a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:13.178174Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 423, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 262.1543330606073, "ttft_ms": 38.075333926826715, "decode_ms": 63.93924995791167}, "total_wall_ms": 27335.403875098564, "ttft_ms": 38.075333926826715, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b71ea1d429262a1a87967978", "attempt_id": "b0e128c653e7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:13.566416Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 434, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 308.5683750687167, "ttft_ms": 93.01670896820724, "decode_ms": 137.62745796702802}, "total_wall_ms": 27510.467209038325, "ttft_ms": 93.01670896820724, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3dae6fbe285117e1df8f2763", "attempt_id": "cb0fb6e83839", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:14.129432Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 426, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 273.71316705830395, "ttft_ms": 36.56879102345556, "decode_ms": 235.06470897700638}, "total_wall_ms": 27516.468874993734, "ttft_ms": 36.56879102345556, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0638f239c24222fa1229afb1", "attempt_id": "2b39e37d8226", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:14.698773Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 273.83679104968905, "ttft_ms": 20.28008399065584, "decode_ms": 161.4671660354361}, "total_wall_ms": 27426.674667047337, "ttft_ms": 20.28008399065584, "input_usage": {"lookups": 0}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f89abde00a3511fcb767c6b8", "attempt_id": "c38158c982f6", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:15.178098Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 426, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 326.53570803813636, "ttft_ms": 142.84400001633912, "decode_ms": 37.35825000330806}, "total_wall_ms": 27477.732083993033, "ttft_ms": 142.84400001633912, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28131c48e5ccee488ad7c979", "attempt_id": "a170dcca9312", "run_group_id": "E4_pilot:A5|closed|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:15.709042Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 428, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 26970.517709036358, "hops": {"1": 26970.517709036358}, "receiver_prefill_ms": 282.45070797856897, "ttft_ms": 22.84904196858406, "decode_ms": 41.67162498924881}, "total_wall_ms": 27318.022292107344, "ttft_ms": 22.84904196858406, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "de37e737ab65b0f6897caefc", "attempt_id": "47ee4aeff51d", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:30.385173Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 387, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 279.1892089881003, "ttft_ms": 36.50570893660188, "decode_ms": 32.469124998897314}, "total_wall_ms": 41615.29637593776, "ttft_ms": 36.50570893660188, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a6aa7f583da4ae3c0e0b0ffe", "attempt_id": "a3f459c8c119", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:30.758498Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 251.707874936983, "ttft_ms": 54.26274996716529, "decode_ms": 607.8502919990569}, "total_wall_ms": 42180.93283311464, "ttft_ms": 54.26274996716529, "input_usage": {"lookups": 0}, "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": "Lot 697-W in La Tuque", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.8000000000000002}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "ad0a49b81076edddfe98ccb0a2aac3fc27477fd4125ee7370c410ad68686bd8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c21ff6fccfa192e57a28126d", "attempt_id": "94d17e46caca", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:31.696131Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 393, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 296.7092500766739, "ttft_ms": 40.478667011484504, "decode_ms": 24.006125051528215}, "total_wall_ms": 41628.5148760071, "ttft_ms": 40.478667011484504, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0e5269c13dc2086fdb6536dc", "attempt_id": "e6c4dc58e856", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:32.082549Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 288.2891249610111, "ttft_ms": 23.196624941192567, "decode_ms": 238.85212501045316}, "total_wall_ms": 41817.867457983084, "ttft_ms": 23.196624941192567, "input_usage": {"lookups": 0}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dc9c975c5bb70631341386eb", "attempt_id": "409f84235c9b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:32.657748Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 439.53279207926244, "ttft_ms": 39.812374976463616, "decode_ms": 39.318916969932616}, "total_wall_ms": 41785.94054200221, "ttft_ms": 39.812374976463616, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "871aad6098ec80c8ef7706a6", "attempt_id": "7d2d3b5114f8", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:33.202763Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 396, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 271.18179202079773, "ttft_ms": 36.562500055879354, "decode_ms": 39.103582967072725}, "total_wall_ms": 41614.36954210512, "ttft_ms": 36.562500055879354, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9ad9da516b604627d3f6c912", "attempt_id": "3193b2a85a69", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:33.579397Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 270.45741595793515, "ttft_ms": 40.00258294399828, "decode_ms": 29.68833304475993}, "total_wall_ms": 41607.32075001579, "ttft_ms": 40.00258294399828, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "715035498ab79a2b81f44c57", "attempt_id": "e0989d7edef0", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:33.947399Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 275.6084579741582, "ttft_ms": 55.9717909200117, "decode_ms": 55.33362505957484}, "total_wall_ms": 41654.20729201287, "ttft_ms": 55.9717909200117, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ed1fcdaf36a53291c9e924a2", "attempt_id": "a4fb1b4a26f7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:34.362025Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 404, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 419.1130839753896, "ttft_ms": 39.30699999909848, "decode_ms": 24.27900000475347}, "total_wall_ms": 41749.896875116974, "ttft_ms": 39.30699999909848, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "73f4af09ac48cb3368efb4f0", "attempt_id": "61c92ca0d85a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:34.870117Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 396, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 277.3915409343317, "ttft_ms": 23.21920800022781, "decode_ms": 40.672958944924176}, "total_wall_ms": 41608.559750020504, "ttft_ms": 23.21920800022781, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ca2bf4434740bc114b62eb16", "attempt_id": "5715d563515a", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:35.236448Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 407, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 293.1877920636907, "ttft_ms": 23.504583979956806, "decode_ms": 19.75833298638463}, "total_wall_ms": 41603.689000126906, "ttft_ms": 23.504583979956806, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f3a6d5322d61530136f83ff9", "attempt_id": "27d5fdf2aafc", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:35.599721Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 399, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 275.4252919694409, "ttft_ms": 50.60212500393391, "decode_ms": 457.95520790852606}, "total_wall_ms": 42051.148542086594, "ttft_ms": 50.60212500393391, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5ceee6cb109a9f06605e0f81", "attempt_id": "61bd15200bb4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:36.409479Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 276.2350409757346, "ttft_ms": 36.925625056028366, "decode_ms": 102.45433310046792}, "total_wall_ms": 41682.82166693825, "ttft_ms": 36.925625056028366, "input_usage": {"lookups": 0}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1a84e67aeb531d89016d5b5a", "attempt_id": "aee31af39367", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:36.850997Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 399, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 253.8094170158729, "ttft_ms": 23.466624901629984, "decode_ms": 63.36649996228516}, "total_wall_ms": 41607.836166978814, "ttft_ms": 23.466624901629984, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "85110bc6b3190f59b49e9b72", "attempt_id": "6f60db916dd1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:41:37.215269Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 401, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 41266.59341703635, "hops": {"1": 26970.517709036358, "2": 14296.07570799999}, "receiver_prefill_ms": 274.76675005164, "ttft_ms": 67.19708407763392, "decode_ms": 128.77075001597404}, "total_wall_ms": 41738.021126016974, "ttft_ms": 67.19708407763392, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "76b6d796336e1e317840fb1c", "attempt_id": "5df96d5283d4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:06.114024Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 387, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 434.24099998082966, "ttft_ms": 39.04991701710969, "decode_ms": 31.205625040456653}, "total_wall_ms": 70165.69816600531, "ttft_ms": 39.04991701710969, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fc94dbf718a58167f824a921", "attempt_id": "dc6605d2c26f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:06.642258Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 272.25916704628617, "ttft_ms": 39.21029099728912, "decode_ms": 27.628041920252144}, "total_wall_ms": 70000.0650001457, "ttft_ms": 39.21029099728912, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3619eddb42017e984d33446c", "attempt_id": "83f40d54d387", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:07.004837Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 393, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 274.0655839443207, "ttft_ms": 21.623375010676682, "decode_ms": 38.64795889239758}, "total_wall_ms": 69995.30962505378, "ttft_ms": 21.623375010676682, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "08efe140f3d8b37ebb3142c7", "attempt_id": "076c182685c3", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:07.361910Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 299.41779107321054, "ttft_ms": 35.41954094544053, "decode_ms": 386.16562506649643}, "total_wall_ms": 70382.03770807013, "ttft_ms": 35.41954094544053, "input_usage": {"lookups": 0}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "46355d5980107b0d50ad7391", "attempt_id": "bb406505e94f", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:08.106166Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 286.68883291538805, "ttft_ms": 38.53062505368143, "decode_ms": 28.296374948695302}, "total_wall_ms": 70014.69154108781, "ttft_ms": 38.53062505368143, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "baee5f9f380888498d172a25", "attempt_id": "61e0e887b002", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:08.483728Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 396, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 277.28187502361834, "ttft_ms": 25.8466670056805, "decode_ms": 36.794958054088056}, "total_wall_ms": 70000.94312510919, "ttft_ms": 25.8466670056805, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3ef3e25f75b54b6323304b3d", "attempt_id": "bb150d88c5d7", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:08.848457Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 252.5524590164423, "ttft_ms": 57.57141602225602, "decode_ms": 44.83908403199166}, "total_wall_ms": 70016.06500009075, "ttft_ms": 57.57141602225602, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "83ccb12f5607138cb98b0f48", "attempt_id": "f857b0ada42e", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:09.226565Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 462.97674998641014, "ttft_ms": 25.975083000957966, "decode_ms": 34.47249997407198}, "total_wall_ms": 70184.42708312068, "ttft_ms": 25.975083000957966, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9f7f6bb80bf68690672b19c5", "attempt_id": "2099e537e308", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:09.774519Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 404, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 291.1054170690477, "ttft_ms": 38.43537496868521, "decode_ms": 25.25058400351554}, "total_wall_ms": 70015.85529209115, "ttft_ms": 38.43537496868521, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a2f6324dd0a27c3ca5940948", "attempt_id": "b848fa50dec4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:10.152988Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 396, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 280.71566694416106, "ttft_ms": 31.540540978312492, "decode_ms": 30.881291953846812}, "total_wall_ms": 70004.17145807296, "ttft_ms": 31.540540978312492, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "76fe93edc90abd2d605e7138", "attempt_id": "f0f43046b195", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:10.520078Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 407, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 280.4425829090178, "ttft_ms": 53.17466694395989, "decode_ms": 65.80154108814895}, "total_wall_ms": 70060.594291077, "ttft_ms": 53.17466694395989, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "3bbcac816ef1680b42a4ad33", "attempt_id": "80b6da5f920b", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:10.946923Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 399, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 426.88737506978214, "ttft_ms": 39.47904089000076, "decode_ms": 251.6760410508141}, "total_wall_ms": 70379.24091599416, "ttft_ms": 39.47904089000076, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "aac8883b41ab7b178522a0bd", "attempt_id": "4be9bc0d99b5", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:11.690811Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 395, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 271.8037909362465, "ttft_ms": 26.052458910271525, "decode_ms": 84.95841606054455}, "total_wall_ms": 70043.9220831031, "ttft_ms": 26.052458910271525, "input_usage": {"lookups": 0}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5d7bd58db9da228985185d62", "attempt_id": "1b2ed4fd77d4", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:12.099184Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 399, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 271.38525003101677, "ttft_ms": 44.309708988294005, "decode_ms": 54.23208302818239}, "total_wall_ms": 70031.21795807965, "ttft_ms": 44.309708988294005, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6ffe28c3db439ceba49c83f2", "attempt_id": "40c054f88ad1", "run_group_id": "E4_pilot:A5|closed|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:12.498995Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "closed", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_closed", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 401, "message_tokens": 305, "message_utf8_bytes": 714, "cache_tensor_bytes": null, "serialized_payload_bytes": 714, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 69660.51195806358, "hops": {"1": 26970.517709036358, "2": 14296.07570799999, "3": 14211.41508303117, "4": 14182.503457996063}, "receiver_prefill_ms": 421.3699590181932, "ttft_ms": 36.40974999871105, "decode_ms": 33.48408394958824}, "total_wall_ms": 70152.49274903908, "ttft_ms": 36.40974999871105, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "3": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}, "4": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 714}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "c888aaee7dfde70b6a019b9c3f4e68f6ae981152d0800f0f484a961ea14ec83a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7e6fea1a59ca5c9fa54dd0b5", "attempt_id": "48036b76e71f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:30.238142Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 380, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 275.42654192075133, "ttft_ms": 35.072874976322055, "decode_ms": 46.134958043694496}, "total_wall_ms": 17575.978875160217, "ttft_ms": 35.072874976322055, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "107fd58c6bdf9b02458c7a51", "attempt_id": "166b81cc0fa3", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:30.620239Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 265.0635000318289, "ttft_ms": 36.58429195638746, "decode_ms": 65.95004198607057}, "total_wall_ms": 17586.971833137795, "ttft_ms": 36.58429195638746, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a09bfb0df683371b495eb3f0", "attempt_id": "3c4aefe74870", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:31.014182Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 386, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 276.90712502226233, "ttft_ms": 54.67020801734179, "decode_ms": 104.78737496305257}, "total_wall_ms": 17655.59470804874, "ttft_ms": 54.67020801734179, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "44fed02dc9251c6fc36067f5", "attempt_id": "6319ebb91248", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:31.473714Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 343.1461669970304, "ttft_ms": 29.80116696562618, "decode_ms": 46.409874921664596}, "total_wall_ms": 17638.694083085284, "ttft_ms": 29.80116696562618, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ab438a8c627d453bd3770585", "attempt_id": "181459b8ccd9", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:31.916204Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 276.74608305096626, "ttft_ms": 41.18958301842213, "decode_ms": 29.934125021100044}, "total_wall_ms": 17567.206042120233, "ttft_ms": 41.18958301842213, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "de31d2e89bb3e962756ef6a4", "attempt_id": "b5f47d8a20f6", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:32.286928Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 389, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 267.33025000430644, "ttft_ms": 41.182000073604286, "decode_ms": 58.70779196266085}, "total_wall_ms": 17586.92329202313, "ttft_ms": 41.182000073604286, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fff41ed039d61b43c81f6712", "attempt_id": "51085a8b86a1", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:32.678013Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 331.1646659858525, "ttft_ms": 137.88504095282406, "decode_ms": 34.031332936137915}, "total_wall_ms": 17722.364292130806, "ttft_ms": 137.88504095282406, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a7cad01892c5da0abde4c99d", "attempt_id": "5fb4573cce10", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:33.205470Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 381, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 274.97366699390113, "ttft_ms": 29.94154195766896, "decode_ms": 37.55395801272243}, "total_wall_ms": 17561.659125145525, "ttft_ms": 29.94154195766896, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "dd14bf89e6966e88bf599c71", "attempt_id": "56cd58871653", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:33.572458Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 397, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 299.23583299387246, "ttft_ms": 19.41091602202505, "decode_ms": 41.20695893652737}, "total_wall_ms": 17578.987167100422, "ttft_ms": 19.41091602202505, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "37a9d30c4fb0c93d6ba413ad", "attempt_id": "fb3ed1d77af5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:33.955693Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 389, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 264.7230840520933, "ttft_ms": 55.738375056535006, "decode_ms": 47.94312501326203}, "total_wall_ms": 17587.650791043416, "ttft_ms": 55.738375056535006, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b25d383f3f68557d45250b28", "attempt_id": "009fda9502ec", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:34.347940Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 400, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 419.0637089777738, "ttft_ms": 39.831040892750025, "decode_ms": 125.07329205982387}, "total_wall_ms": 17803.23775007855, "ttft_ms": 39.831040892750025, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "cfecc99ff9c2c29e5d1b0bbd", "attempt_id": "091d3785a4ce", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:34.957080Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 392, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 289.53554201871157, "ttft_ms": 39.23549992032349, "decode_ms": 24.60791601333767}, "total_wall_ms": 17572.798042092472, "ttft_ms": 39.23549992032349, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8eeb6712740872d2bbe509bd", "attempt_id": "442bf7a14e10", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:35.334740Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 388, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 261.04945805855095, "ttft_ms": 32.085124985314906, "decode_ms": 52.175666904076934}, "total_wall_ms": 17564.668791019358, "ttft_ms": 32.085124985314906, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b2f4a02726ccc88185c84b61", "attempt_id": "60e3fb2f952b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:35.703690Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 392, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 313.9067080337554, "ttft_ms": 101.87125008087605, "decode_ms": 137.34904094599187}, "total_wall_ms": 17772.66270807013, "ttft_ms": 101.87125008087605, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4287c15cb6f6d9d57d2481de", "attempt_id": "db941f71daba", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:36.283509Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 394, "message_tokens": 299, "message_utf8_bytes": 1621, "cache_tensor_bytes": null, "serialized_payload_bytes": 1621, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 17218.6268750811, "hops": {"1": 17218.6268750811}, "receiver_prefill_ms": 276.88112494070083, "ttft_ms": 38.352209026925266, "decode_ms": 21.807334036566317}, "total_wall_ms": 17556.449791183695, "ttft_ms": 38.352209026925266, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "df2941d96846f81eb0594f50e7e7c5e356f11017258161affd0cb03f67c5f42b", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "b00dc4f70027928b086244fe", "attempt_id": "bebe415163d7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:46.987457Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 311, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 248.2379999710247, "ttft_ms": 24.14783404674381, "decode_ms": 38.86595799122006}, "total_wall_ms": 27869.754709186964, "ttft_ms": 24.14783404674381, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "63710144d0dbdaaf295a8670", "attempt_id": "e331aa072b75", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:47.324445Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 227.19408408738673, "ttft_ms": 19.72741703502834, "decode_ms": 40.644625085406005}, "total_wall_ms": 27846.199834137224, "ttft_ms": 19.72741703502834, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9770a705f1a4ed625d276ffe", "attempt_id": "eb5e45ce14fc", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:47.637228Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 317, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 265.7021669438109, "ttft_ms": 42.024333961308, "decode_ms": 111.0911249415949}, "total_wall_ms": 27977.277876110747, "ttft_ms": 42.024333961308, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4040f20b6a3807b4540b5ee2", "attempt_id": "603b4be9b03f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:48.079570Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 325.0352500472218, "ttft_ms": 40.16895801760256, "decode_ms": 38.73862500768155}, "total_wall_ms": 27962.34450105112, "ttft_ms": 40.16895801760256, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6d461b12483ce9bdace3b0fc", "attempt_id": "8469acc4000e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:48.508816Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 236.85529199428856, "ttft_ms": 38.281042012386024, "decode_ms": 24.49829096440226}, "total_wall_ms": 27857.973292120732, "ttft_ms": 38.281042012386024, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "276f678f3abf8f6b98adb2ec", "attempt_id": "9287bb0c6312", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:48.831782Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 320, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 220.75837501324713, "ttft_ms": 16.819249955005944, "decode_ms": 49.18129194993526}, "total_wall_ms": 27845.176792121492, "ttft_ms": 16.819249955005944, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ebe650bbc104adcd1528b1a8", "attempt_id": "7f42443654bb", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:49.141503Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 275.42954101227224, "ttft_ms": 43.349332991056144, "decode_ms": 96.6075409669429}, "total_wall_ms": 27974.102750071324, "ttft_ms": 43.349332991056144, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "69020739572a5d3cb96ecc11", "attempt_id": "734bc2b019f7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:49.583329Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 312.55324999801815, "ttft_ms": 31.41162497922778, "decode_ms": 40.455249953083694}, "total_wall_ms": 27942.875542095862, "ttft_ms": 31.41162497922778, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "712b6f1781e7ea79c85b8b12", "attempt_id": "c06332b84add", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:49.995211Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 328, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 258.8493329240009, "ttft_ms": 38.654582924209535, "decode_ms": 23.169292020611465}, "total_wall_ms": 27879.089084104635, "ttft_ms": 38.654582924209535, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "5a1799f88d82cc327dfeba46", "attempt_id": "12da2cff9593", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:50.340700Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 320, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 214.9470419390127, "ttft_ms": 19.287790986709297, "decode_ms": 47.35008301213384}, "total_wall_ms": 27840.03512514755, "ttft_ms": 19.287790986709297, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "18cf8a842bba8b7e235377d5", "attempt_id": "cbabba4fc587", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:50.646425Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 331, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 256.5862920600921, "ttft_ms": 61.8691670242697, "decode_ms": 331.63979195524007}, "total_wall_ms": 28208.65170902107, "ttft_ms": 61.8691670242697, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "451cca5e10fd5bb3f4fcc095", "attempt_id": "a350f4e02e5e", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:51.320486Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 323, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 247.9824589099735, "ttft_ms": 41.55579104553908, "decode_ms": 31.24312497675419}, "total_wall_ms": 27879.28629212547, "ttft_ms": 41.55579104553908, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fe83e20e823803f6a3d7d2d5", "attempt_id": "c023e9310110", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:51.666351Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 319, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 231.61975003313273, "ttft_ms": 30.16633295919746, "decode_ms": 30.40895808953792}, "total_wall_ms": 27850.6170011824, "ttft_ms": 30.16633295919746, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "923831205dcbfda683e668d5", "attempt_id": "6aa80cb8394d", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:51.983346Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 323, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 230.88045802433044, "ttft_ms": 39.816917036660016, "decode_ms": 60.85179094225168}, "total_wall_ms": 27890.325542190112, "ttft_ms": 39.816917036660016, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "9e5b0c14669830e8e9edacc8", "attempt_id": "63f154e1300f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:42:52.339807Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 325, "message_tokens": 230, "message_utf8_bytes": 1294, "cache_tensor_bytes": null, "serialized_payload_bytes": 1294, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27557.806917116977, "hops": {"1": 17218.6268750811, "2": 10339.180042035878}, "receiver_prefill_ms": 322.84299994353205, "ttft_ms": 116.37683294247836, "decode_ms": 30.620375066064298}, "total_wall_ms": 28028.28800014686, "ttft_ms": 116.37683294247836, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "72e415464c6f6d244984d880ff6ea004de1a85901891302e8944a72bd7d02f48", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "13fba0204824f0442a3ddb3c", "attempt_id": "36063202af93", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:11.970944Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 298, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 283.39908295311034, "ttft_ms": 137.2937080450356, "decode_ms": 40.63770896755159}, "total_wall_ms": 47151.65679214988, "ttft_ms": 137.2937080450356, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4a3214c3d5b40cce76dc2a57", "attempt_id": "a777b3c48a89", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:12.460386Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 253.53912508580834, "ttft_ms": 42.41641703993082, "decode_ms": 37.700625020079315}, "total_wall_ms": 47024.1366251139, "ttft_ms": 42.41641703993082, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "bd6acc5729483549f78d330f", "attempt_id": "e93bca8bc7d2", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:12.819026Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 304, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 231.41541599761695, "ttft_ms": 25.1672500744462, "decode_ms": 38.85799995623529}, "total_wall_ms": 46985.78875011299, "ttft_ms": 25.1672500744462, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e8bfafeebad8546cd4734d62", "attempt_id": "d5d566ed70a5", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:13.141012Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 211.10337495338172, "ttft_ms": 61.56704109162092, "decode_ms": 37.320791045203805}, "total_wall_ms": 47000.537625164725, "ttft_ms": 61.56704109162092, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2867b70dc8246438a63e6d42", "attempt_id": "6fba236c6344", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:13.476487Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 425.003582960926, "ttft_ms": 23.862750036641955, "decode_ms": 42.85850003361702}, "total_wall_ms": 47181.97666713968, "ttft_ms": 23.862750036641955, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1ef51bcb1ec54d216c4530a3", "attempt_id": "e970d618ce9f", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:13.991528Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 307, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 244.14570804219693, "ttft_ms": 30.506124952808022, "decode_ms": 40.026500006206334}, "total_wall_ms": 47004.9285000423, "ttft_ms": 30.506124952808022, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "008ec1782d6cea010c585eae", "attempt_id": "75c7689026cd", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:14.329217Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 242.4776660045609, "ttft_ms": 22.901542019098997, "decode_ms": 35.27700004633516}, "total_wall_ms": 46991.055750171654, "ttft_ms": 22.901542019098997, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1758b9257ed478c5d0bd9dcb", "attempt_id": "bccb25200db7", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:14.653573Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 299, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 243.36858303286135, "ttft_ms": 54.9756670370698, "decode_ms": 55.61137502081692}, "total_wall_ms": 47044.40562508535, "ttft_ms": 54.9756670370698, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "fd6376b818b18c85f96abcca", "attempt_id": "90f2677a5f2b", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:15.033679Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 315, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 396.70508401468396, "ttft_ms": 21.494542015716434, "decode_ms": 43.95262501202524}, "total_wall_ms": 47152.60566712823, "ttft_ms": 21.494542015716434, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2a63b0d82cf3f3981388001e", "attempt_id": "e0b5f8439606", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:15.520589Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 307, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 249.39545802772045, "ttft_ms": 21.36895793955773, "decode_ms": 42.84070909488946}, "total_wall_ms": 47004.04495908879, "ttft_ms": 21.36895793955773, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e264ad3e1f1e8a0ebb250e05", "attempt_id": "6cd4517f18a0", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:15.859060Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 318, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 240.3853329597041, "ttft_ms": 23.135125054977834, "decode_ms": 161.60383401438594}, "total_wall_ms": 47115.72441610042, "ttft_ms": 23.135125054977834, "input_usage": {"lookups": 0}, "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": "Ixruk-6", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "a4dadc61e6fa508ecedb4a8a4b025cce77334c45b5b0db3336e15fed6a121d18"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "ee1b0836f986d3eecebbe00b", "attempt_id": "4a136f559911", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:16.312505Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 310, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 325.02470805775374, "ttft_ms": 142.13687495794147, "decode_ms": 46.10612499527633}, "total_wall_ms": 47203.68295814842, "ttft_ms": 142.13687495794147, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8fa20468f8572dc061504482", "attempt_id": "0106ad9e9378", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:16.851189Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 306, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 242.17637500260025, "ttft_ms": 43.71504101436585, "decode_ms": 36.34095808956772}, "total_wall_ms": 47012.45141716208, "ttft_ms": 43.71504101436585, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c5d1e515c3c0e24ae6837918", "attempt_id": "8c26c0355ffa", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:17.197107Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 310, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 233.4276249166578, "ttft_ms": 24.890333064831793, "decode_ms": 34.94608402252197}, "total_wall_ms": 46983.44062513206, "ttft_ms": 24.890333064831793, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c039c794962252147b1ec0d", "attempt_id": "c8089b968166", "run_group_id": "E4_pilot:A4|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:17.513181Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A4", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 312, "message_tokens": 217, "message_utf8_bytes": 1224, "cache_tensor_bytes": null, "serialized_payload_bytes": 1224, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 46689.74075012375, "hops": {"1": 17218.6268750811, "2": 10339.180042035878, "3": 9600.960624986328, "4": 9530.973208020441}, "receiver_prefill_ms": 211.08912501949817, "ttft_ms": 64.6958340657875, "decode_ms": 59.181082993745804}, "total_wall_ms": 47025.230791186914, "ttft_ms": 64.6958340657875, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1621}, "2": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1294}, "3": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}, "4": {"facts_total": 62, "facts_retained": 8, "retention_rate": 0.12903225806451613, "retention_by_kind": {"person": 0.0, "date": 0.0, "amount": 0.0, "status": 0.6666666666666666, "place": 0.0, "identifier": 0.0, "quantity": 0.0}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 1224}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "cc31f4882fecb841f505a224fcb13f7fc6ef41e83aaa8abcca2e152e97bb482a", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "160eebb661f1b77086d2b993", "attempt_id": "f91a2a1e26d7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:44.970423Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 414, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 270.6475410377607, "ttft_ms": 60.88866607751697, "decode_ms": 59.450834058225155}, "total_wall_ms": 27488.37937496137, "ttft_ms": 60.88866607751697, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c5367bd8397d70c14a561043", "attempt_id": "629d96073597", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:45.388044Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 279.9176670378074, "ttft_ms": 94.74400000181049, "decode_ms": 457.78129203245044}, "total_wall_ms": 27929.76683296729, "ttft_ms": 94.74400000181049, "input_usage": {"lookups": 0}, "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": "Lot 697-W in La Tuque", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.8000000000000002}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "ad0a49b81076edddfe98ccb0a2aac3fc27477fd4125ee7370c410ad68686bd8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "05dd733de1abbfaf3e3afd34", "attempt_id": "3495e636368f", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:46.244671Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 420, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 295.0278330827132, "ttft_ms": 25.95345804002136, "decode_ms": 221.60412499215454}, "total_wall_ms": 27639.817166840658, "ttft_ms": 25.95345804002136, "input_usage": {"lookups": 0}, "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": "2.7230", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "7e908e02eb04c1aa8a80abf6fdd1511a7fa4f87b430c0f0452f47b4063b159f0"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7bff8414107ec36687966327", "attempt_id": "3e32035f8bb4", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:46.812070Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 287.26516698952764, "ttft_ms": 82.87679101340473, "decode_ms": 320.31479198485613}, "total_wall_ms": 27787.688790936954, "ttft_ms": 82.87679101340473, "input_usage": {"lookups": 0}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "66bf801fdbc0572b5771290d", "attempt_id": "6fe5838c69c3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:47.527650Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 280.8273750124499, "ttft_ms": 25.43437492568046, "decode_ms": 40.728542022407055}, "total_wall_ms": 27444.343749899417, "ttft_ms": 25.43437492568046, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "182c8688403815778936de71", "attempt_id": "4bb97a8fb247", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:47.899768Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 423, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 290.25266703683883, "ttft_ms": 20.530459005385637, "decode_ms": 19.880249979905784}, "total_wall_ms": 27427.920332876965, "ttft_ms": 20.530459005385637, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "28eb223f8f751a603c6710d1", "attempt_id": "2cd1b4ea08b8", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:48.257055Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 313.13666701316833, "ttft_ms": 48.74133400153369, "decode_ms": 39.551666006445885}, "total_wall_ms": 27498.79604089074, "ttft_ms": 48.74133400153369, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "no", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "9390298f3fb0c5b160498935d79cb139aef28e1c47358b4bbba61862b9c26e59"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f40adce510f982728c4508ae", "attempt_id": "5878277f3614", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:48.684693Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 415, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 430.48591702245176, "ttft_ms": 34.639458055607975, "decode_ms": 44.426333042792976}, "total_wall_ms": 27606.746791978367, "ttft_ms": 34.639458055607975, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a8c6cf5df2f3b0bea50da2ee", "attempt_id": "fd5bfba4d982", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:49.219042Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 431, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 281.81141696404666, "ttft_ms": 37.18650003429502, "decode_ms": 26.164958020672202}, "total_wall_ms": 27442.399916006252, "ttft_ms": 37.18650003429502, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "6903e58eaa483cf04e9a4870", "attempt_id": "7cb27c8f82b7", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:49.587549Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 423, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 268.14291602931917, "ttft_ms": 58.78108402248472, "decode_ms": 56.62470799870789}, "total_wall_ms": 27480.773583054543, "ttft_ms": 58.78108402248472, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "689ca339784ff779b2ecd86e", "attempt_id": "506a7a40b5d2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:49.994737Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 434, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 288.11983310151845, "ttft_ms": 141.8992499820888, "decode_ms": 98.84162503294647}, "total_wall_ms": 27626.158999977633, "ttft_ms": 141.8992499820888, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d1329e3e7020581036c86edf", "attempt_id": "967191a9b0f0", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:50.548341Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 426, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 287.3843329725787, "ttft_ms": 26.010749978013337, "decode_ms": 272.15291606262326}, "total_wall_ms": 27682.766207959503, "ttft_ms": 26.010749978013337, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "33f449f87089733c483bbc4c", "attempt_id": "b00adf0736db", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:51.156843Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 422, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 276.3070419896394, "ttft_ms": 39.81787501834333, "decode_ms": 159.83083297032863}, "total_wall_ms": 27573.238041019067, "ttft_ms": 39.81787501834333, "input_usage": {"lookups": 0}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c3e14a6c375729dddbb82627", "attempt_id": "49d414142656", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:51.657184Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 426, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 416.62649996578693, "ttft_ms": 36.41037503257394, "decode_ms": 28.33929203916341}, "total_wall_ms": 27578.631582902744, "ttft_ms": 36.41037503257394, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0c9acb472b7fca809a1cadc4", "attempt_id": "5ecf062a29ec", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h1", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:43:52.164726Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 1, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 428, "message_tokens": 332, "message_utf8_bytes": 798, "cache_tensor_bytes": null, "serialized_payload_bytes": 798, "side_information_bytes": 0, "fetched_bytes": 0, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 27096.659707953222, "hops": {"1": 27096.659707953222}, "receiver_prefill_ms": 278.58679194469005, "ttft_ms": 26.868875022046268, "decode_ms": 38.55549998115748}, "total_wall_ms": 27441.26074993983, "ttft_ms": 26.868875022046268, "input_usage": {"lookups": 0}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "9fdbd56a92e56663f9bd6bef1f5cfd307786461e33ee1df952e0946e089385a2", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "097fb6bbc6e11ae2a0b43487", "attempt_id": "74294f728e25", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:05.340505Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 351, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 279.93191697169095, "ttft_ms": 39.8836659733206, "decode_ms": 24.67970794532448}, "total_wall_ms": 40247.69929097965, "ttft_ms": 39.8836659733206, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "88c358eca4f47e2dbd68d220", "attempt_id": "ce108620b0ff", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:05.711381Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 256.9607909535989, "ttft_ms": 27.720832964405417, "decode_ms": 582.2112089954317}, "total_wall_ms": 40769.948415923864, "ttft_ms": 27.720832964405417, "input_usage": {"lookups": 2}, "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": "Lot 697-W in La Tuque", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.8000000000000002}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "ad0a49b81076edddfe98ccb0a2aac3fc27477fd4125ee7370c410ad68686bd8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "396f7d526fad80f3542d152f", "attempt_id": "7fcee1c7f602", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:06.603047Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 357, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 314.2585839377716, "ttft_ms": 32.1073749801144, "decode_ms": 33.036790904589}, "total_wall_ms": 40282.50029194169, "ttft_ms": 32.1073749801144, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "1c3bc32fffb44fa3f2f68a12", "attempt_id": "0befcf947f72", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:07.006001Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 264.8465409874916, "ttft_ms": 41.16704105399549, "decode_ms": 150.53554100450128}, "total_wall_ms": 40359.614457935095, "ttft_ms": 41.16704105399549, "input_usage": {"lookups": 2}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "440ebb5afae25e5f484ee161", "attempt_id": "a5249c497dc3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:07.486700Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 267.4254580633715, "ttft_ms": 56.814791983924806, "decode_ms": 47.14291705749929}, "total_wall_ms": 40274.38283304218, "ttft_ms": 56.814791983924806, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "4e6df94f30e34d3fb2e29935", "attempt_id": "74e375dcfd92", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:07.883251Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 360, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 405.3987079532817, "ttft_ms": 33.97350001614541, "decode_ms": 33.5926249390468}, "total_wall_ms": 40376.01787492167, "ttft_ms": 33.97350001614541, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "a7c2d93e322778b4c53768d7", "attempt_id": "9ecee07abf83", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:08.381782Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 267.006375011988, "ttft_ms": 39.31845899205655, "decode_ms": 23.841166985221207}, "total_wall_ms": 40233.17866702564, "ttft_ms": 39.31845899205655, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d54616b9fe68bd4da0717932", "attempt_id": "6a27e737cddc", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:08.736818Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 259.36437502969056, "ttft_ms": 20.42029204312712, "decode_ms": 19.512583035975695}, "total_wall_ms": 40202.327499049716, "ttft_ms": 20.42029204312712, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "36ca03a47b747c3b1528ebb5", "attempt_id": "9ee02387d382", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:09.060945Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 368, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 292.2667919192463, "ttft_ms": 41.36333300266415, "decode_ms": 62.52862501423806}, "total_wall_ms": 40299.24412397668, "ttft_ms": 41.36333300266415, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "27e830ca543e75786437627a", "attempt_id": "035b0d57d187", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:09.482759Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 360, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 399.99179099686444, "ttft_ms": 23.801833973266184, "decode_ms": 39.44424993824214}, "total_wall_ms": 40366.29741592333, "ttft_ms": 23.801833973266184, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "99c9ea587f8ca9afa96bef6d", "attempt_id": "7be24c70d779", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:09.969998Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 371, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 279.2944998946041, "ttft_ms": 44.55966700334102, "decode_ms": 23.039666935801506}, "total_wall_ms": 40249.87812503241, "ttft_ms": 44.55966700334102, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e9545555a1d6898046c0a07c", "attempt_id": "1ddddf2923ee", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:10.340271Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 363, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 242.69366706721485, "ttft_ms": 42.33170789666474, "decode_ms": 531.8588750669733}, "total_wall_ms": 40719.87962501589, "ttft_ms": 42.33170789666474, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8d208840edc36768a4a9e169", "attempt_id": "25677d93dbfe", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:11.180573Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 255.0337080610916, "ttft_ms": 38.70770800858736, "decode_ms": 86.45745902322233}, "total_wall_ms": 40283.168916008435, "ttft_ms": 38.70770800858736, "input_usage": {"lookups": 2}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "87ba2816f845b58bc6c95d79", "attempt_id": "9c5d8eb11e19", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:11.584052Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 363, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 254.396167001687, "ttft_ms": 36.791041027754545, "decode_ms": 37.05366689246148}, "total_wall_ms": 40231.25524900388, "ttft_ms": 36.791041027754545, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "143bffea3fc43631be11becd", "attempt_id": "7e0c98118d0b", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h2", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:11.935378Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 2, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 365, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 956, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 39902.49670797493, "hops": {"1": 27096.659707953222, "2": 12805.837000021711}, "receiver_prefill_ms": 269.3365829763934, "ttft_ms": 53.671125089749694, "decode_ms": 59.70070802140981}, "total_wall_ms": 40285.86991596967, "ttft_ms": 53.671125089749694, "input_usage": {"lookups": 2}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "21efb812d1c3fcacb8994c01", "attempt_id": "342c30493378", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:36.861667Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q00_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 351, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 234.6168749500066, "ttft_ms": 56.990375043824315, "decode_ms": 41.64466599468142}, "total_wall_ms": 64751.01620785426, "ttft_ms": 56.990375043824315, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f619562a6855f24b36f74ef8", "attempt_id": "7d494a8f9e85", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:37.219738Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q01_single_fact", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 447.6237919880077, "ttft_ms": 26.34450001642108, "decode_ms": 352.4934999877587}, "total_wall_ms": 65244.137873873115, "ttft_ms": 26.34450001642108, "input_usage": {"lookups": 4}, "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": "Lot 697-W in La Tuque", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 0.0, "abstained": 0.0, "f1": 0.8000000000000002}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "ad0a49b81076edddfe98ccb0a2aac3fc27477fd4125ee7370c410ad68686bd8d"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "d208e3a0649999cfc1eeb0c9", "attempt_id": "7811a2a1b747", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:38.072135Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q02_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 357, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 251.57420802861452, "ttft_ms": 30.44829098507762, "decode_ms": 31.680625048466027}, "total_wall_ms": 64731.04224994313, "ttft_ms": 30.44829098507762, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "e112a3c4b3bf0146323c1b7f", "attempt_id": "3e74ff4163a2", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:38.411321Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q03_exact_copy", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 266.7225830955431, "ttft_ms": 66.40237499959767, "decode_ms": 372.31045903172344}, "total_wall_ms": 65122.8469988564, "ttft_ms": 66.40237499959767, "input_usage": {"lookups": 4}, "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": "Dagny Mbeki", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_8000.json", "task_scores": {"correct": 1.0, "abstained": 0.0, "f1": 1.0}, "status": "success", "error_code": null, "retry_of": null, "log_artifact": "runs/logs/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "e17363b2d01a8f3b3d1f3d929af2daab67bfe581aa17c9b4c1c338747a33eff3"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "7620a2d1cbefdf26b75fcd24", "attempt_id": "c77a5a4647b6", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:39.142422Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q04_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 256.97016704361886, "ttft_ms": 37.18133398797363, "decode_ms": 27.326166979037225}, "total_wall_ms": 64739.02370687574, "ttft_ms": 37.18133398797363, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8330ffd962fa3eae05d4ac96", "attempt_id": "3b8c76fe7cf9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:39.489686Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q05_multi_hop", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 360, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 259.1539160348475, "ttft_ms": 25.248249992728233, "decode_ms": 36.921125021763146}, "total_wall_ms": 64738.685623859055, "ttft_ms": 25.248249992728233, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "38b15c678fd11fc7f2438fd9", "attempt_id": "b11fa3d90362", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:39.835138Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q06_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 268.0318750208244, "ttft_ms": 59.76712505798787, "decode_ms": 63.17604100331664}, "total_wall_ms": 64808.405831921846, "ttft_ms": 59.76712505798787, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "c35a247330735b9192153c63", "attempt_id": "4b4374b8069c", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:40.249549Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q07_negation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 352, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 402.84145798068494, "ttft_ms": 35.810249974019825, "decode_ms": 28.01474998705089}, "total_wall_ms": 64884.03662387282, "ttft_ms": 35.810249974019825, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f35dbae83e8ef94b5fa6f589", "attempt_id": "f5c38053330a", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:40.740548Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q08_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 368, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 257.12954194750637, "ttft_ms": 21.95612504146993, "decode_ms": 43.78654097672552}, "total_wall_ms": 64740.26341584977, "ttft_ms": 21.95612504146993, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "f8ff2683423ff4fa279d60a4", "attempt_id": "f62d058b0bd9", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:41.086465Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q09_revision", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 360, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 274.0311670349911, "ttft_ms": 37.08729206118733, "decode_ms": 30.249833944253623}, "total_wall_ms": 64758.73174984008, "ttft_ms": 37.08729206118733, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "20b91d38efabc52634a0a2d6", "attempt_id": "2a372e112de3", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:41.452018Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q10_temporal", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 371, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 261.7311250651255, "ttft_ms": 58.635500026866794, "decode_ms": 55.70195789914578}, "total_wall_ms": 64793.572582886554, "ttft_ms": 58.635500026866794, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "2779b5279d9bb792cb6711c1", "attempt_id": "e1cb21c6bea4", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:41.853264Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q11_aggregation", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 363, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 402.6515829609707, "ttft_ms": 42.58762497920543, "decode_ms": 247.06700001843274}, "total_wall_ms": 65109.66579092201, "ttft_ms": 42.58762497920543, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 7}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "3240125", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "40e371f429efb209db956af4c4d9bfaa16f93bc371169df19625528850631d70"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "74ce15ada7f83eed6fe0da4f", "attempt_id": "0e1c4b71df60", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:42.568987Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q12_unit_conversion", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 359, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 265.168791054748, "ttft_ms": 35.4367500403896, "decode_ms": 73.59254197217524}, "total_wall_ms": 64791.59012378659, "ttft_ms": 35.4367500403896, "input_usage": {"lookups": 4}, "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": "0.8", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "1e9d7c27c8bbc8ddf0055c93e064a62fa995d177fee28cc8fa949bc8a4db06f4"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "8c6c184d73567ec456abde64", "attempt_id": "3f36c8008c56", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:42.967883Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q13_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 363, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 269.36033403035253, "ttft_ms": 59.82704204507172, "decode_ms": 65.912791993469}, "total_wall_ms": 64812.51579080708, "ttft_ms": 59.82704204507172, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}
{"schema_version": "wp30-attempt-1", "study_id": "WP30", "experiment_id": "E4_pilot", "trial_id": "0cf743571a2b8090078a6f68", "attempt_id": "019610e201ff", "run_group_id": "E4_pilot:A5|repairable|qwen3b:h4", "protocol_hash": "unfrozen", "config_hash": "2b706fe0dd1e4bbc04fbc96c8deb9109308855fdc080eeb3fc11e59cb1d91751", "git_commit": null, "timestamp_utc": "2026-10-30T04:44:43.389160Z", "task_family": "synthetic", "document_id": "syn_02007_8000", "document_hash": "da6611bb7f87fd8656ea7643488613a1e9dec85b7bc0c29b9f052dcde1b89142", "question_id": "q14_unanswerable", "split": "dev", "cluster_id": "syn_02007_8000", "sender_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "receiver_model_revision": "mlx-community/Qwen2.5-3B-Instruct-bf16@bec1ad658e380044bbfe06d4a2268801d3379a11", "tokenizer_hashes": {"receiver": "5c5a439ba2677fe28e1dea9b22371b270ecd5e34acbc849bc619689605f9547b"}, "backend_versions": {"mlx": "0.32.4", "mlx_lm": "0.32.2"}, "hardware_id": "laptop-m5max-48g", "device_state": {}, "weight_precision": "bf16", "method_id": "A5", "implementation_fidelity": "custom", "compression_config": {"mode": "repairable", "agents": ["qwen3b"], "budget_tokens": 512}, "question_access": "blind", "topology": "chain_repairable", "handoff_index": 4, "reuse_count": 15, "cache_hit": false, "cache_compatibility_hash": null, "random_seed": 0, "source_tokens_sender": null, "source_tokens_receiver": 5845, "query_tokens": 365, "message_tokens": 270, "message_utf8_bytes": 641, "cache_tensor_bytes": null, "serialized_payload_bytes": 641, "side_information_bytes": 0, "fetched_bytes": 1802, "transport_bytes_if_measured": null, "timing_spans": {"chain_ms": 64416.88766586594, "hops": {"1": 27096.659707953222, "2": 12805.837000021711, "3": 12302.034957916476, "4": 12212.355999974534}, "receiver_prefill_ms": 401.0318749351427, "ttft_ms": 24.29620793554932, "decode_ms": 39.2453329404816}, "total_wall_ms": 64882.01045792084, "ttft_ms": 24.29620793554932, "input_usage": {"lookups": 4}, "output_usage": {"output_tokens": 1}, "provider_cached_usage": null, "process_memory": null, "backend_memory": null, "peak_prefill_memory": null, "peak_decode_memory": null, "mapper_memory": null, "cost_with_source": null, "prediction_artifact": "unknown", "evidence_artifact": {"ledger": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "ledger_by_hop": {"1": {"facts_total": 62, "facts_retained": 7, "retention_rate": 0.11290322580645161, "retention_by_kind": {"person": 0.3333333333333333, "date": 0.0, "amount": 0.08333333333333333, "status": 0.0, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 2, "unsupported_identifiers": 0, "distorted_sentences": 1, "message_bytes": 798}, "2": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "3": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}, "4": {"facts_total": 62, "facts_retained": 10, "retention_rate": 0.16129032258064516, "retention_by_kind": {"person": 0.16666666666666666, "date": 0.0, "amount": 0.08333333333333333, "status": 0.3333333333333333, "place": 0.16666666666666666, "identifier": 0.16666666666666666, "quantity": 0.16666666666666666}, "unsupported_numbers": 0, "unsupported_identifiers": 0, "distorted_sentences": 0, "message_bytes": 641}}}, "reference_key_id": "keys/synthetic/syn_02007_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/E4_pilot.log", "payload_hash": "3a9ef82b1db694d77e2fedc026bb31176dcfc6e4899f10c4469b2634de610819", "output_hash": "b23a6a8439c0dde5515893e7c90c1e3233b8616e634470f20dc4928bcf3609bc"}