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%
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