SPB Git

spb/modelmap Public License

Internal cartography of local LLMs on Apple Silicon — registered, gated, negative-first. Public atlas at modelmap.io.

Python 66.3% JavaScript 24.5% CSS 8.1% Shell 0.7%
3.0 KB · 77 lines json
Raw Blame History
1{2  "map_id": "atlas/qwen3-0.6b-4bit/probes/v2",3  "map_type": "probes",4  "model_id": "mlx-community/Qwen3-0.6B-4bit",5  "model_hash": "392e8d466d56100ada00eb82031fb854297fc9e389b7d303eba3af114e87bce2",6  "quantization": "q4 (mlx)",7  "commit": "1dcd820126ca96fe99c1b473947f3c7e54038500",8  "config": "results/expA_probe_reliability/20260812T063856Z/results.json",9  "created": "2026-08-12",10  "hardware_manifest": {11    "author": "Simon-Pierre Boucher",12    "contact": "contact@spboucher.ai",13    "website": "https://modelmap.io",14    "chip": {15      "brand": "Apple M5 Max",16      "cores_total": 18,17      "cores_performance": 6,18      "cores_efficiency": 1219    },20    "memory": {21      "unified_gb": 48.0,22      "pagesize": 1638423    },24    "os": {25      "system": "Darwin",26      "version": "27.0",27      "arch": "arm64"28    },29    "software": {30      "python": "3.14.4",31      "numpy": "2.5.2",32      "mlx": "0.32.0",33      "torch": "2.13.0",34      "safetensors": "0.8.0"35    }36  },37  "confidence_level": 1,38  "regenerate_command": ".venv/bin/python benchmarks/promptsets/make_promptsets_v2.py && .venv/bin/python experiments/micro/expA_probe_reliability/implementation/benchmark_v2.py && .venv/bin/python experiments/micro/expA_probe_reliability/implementation/make_mapcard_v2.py",39  "seeds": [40    0,41    1,42    2,43    3,44    445  ],46  "prompt_sets": [47    "agreement_A.jsonl#622c5e0966d2b8e8",48    "agreement_B.jsonl#243b85e2c207be6c",49    "arith_valid_A.jsonl#a4bb4c645b3f1797",50    "arith_valid_B.jsonl#adc68ef876a823ec",51    "word_order_A.jsonl#9052b924930aa5cd",52    "word_order_B.jsonl#3012db3b5053ba5d"53  ],54  "controls": [55    "shuffled-label (every probe)",56    "random-init architecture twin",57    "BH-FDR q=0.05",58    "v1 positive control (ceiling check)",59    "class token-overlap certificates in promptset manifest"60  ],61  "methods_in_agreement": [],62  "interventions": [63    "layer-skip ablation (expC run #1, 2026-08-12): top-5 differential layers NOT confirmed — damage below random-5 mean; see experiments/micro/expC_causal_verification/analysis.md"64  ],65  "replication_rate": 0.5366,66  "per_dataset_agreement": null,67  "ablation_schemes": [],68  "featurizer_class": "natural-basis (mean-pooled + last-token residual)",69  "intervention_protocol": "none (observational map — Level 1 by design)",70  "negative_result": false,71  "notes": "DIFFERENTIAL map (real minus random-init twin), per the doctrine adopted after v1. Mixed outcome by property: agreement and arith_valid carry trained-model signal above the architecture prior; word_order is null-dominated and flagged as such. Strict twin gate (<0.05) still fails on word_order/agreement — only differential claims are published. CAUSAL CHECK: expC run #1 layer-skip ablation did NOT confirm the top differential layers (survival 0/1); map remains Level 1 and its layer ranking must not be read as causal.",72  "author": "Simon-Pierre Boucher",73  "contact": "contact@spboucher.ai",74  "website": "https://modelmap.io",75  "schema_version": "0.1"76}77