{
  "schema": "ilxyr.experiment_evidence_summary.v1",
  "public_id": "EXP-008",
  "experiment_id": "solomon.q22-compositional-routing.three-seed.v1",
  "shared_task_id": "zero-solomon.q22-compositional-routing.v1",
  "resolved_outcome": "no_go",
  "registration": {
    "repository": "https://github.com/cenetex/ilXyr.git",
    "experiment_commit": "033644f9c555c21c000b3affb034b363ed28430c",
    "registration_merge_commit": "8168c86a2996b731d703f662537432e4cf9d2a02",
    "shared_task_ref": "artifact://sha256/97236b3f02109ab9fd64741304035164e0cd60f999e4b9f1851abb17006c9774",
    "harness_sha256": "e3f84e9dbc3b5a2223f554aeb2966750d8df74e4148627a1da3f0d8f787e969c",
    "merged_before_evaluation": true
  },
  "datasets": {
    "training_records": 10000,
    "training_sha256": "1f997d174ea1e3f3675612e24ecced15f8145abd29819e532b2f890f36245e8f",
    "encoded_training_sha256": "6fc74419d063ecbba1bf86bcb5e9cc0df8e134fc3b401ea67e48be4edf2a99a5",
    "promotion_records": 1000,
    "promotion_sha256": "42902d61d9edc2ceb5a33875deddbdd9e9b523b06fb1aa13a22249acbf2983fe",
    "blind_inputs_sha256": "871ab80d0b1fc55df785fd33f53baf1395efb2dd0db228aa330e6e5eef29ee45",
    "prefix_only_exact_rate_ppm": 200000,
    "train_template_families": 30,
    "promotion_template_families": 20
  },
  "source": {
    "repository": "https://github.com/atimics/nsrl.git",
    "zero_dataset_commit": "8deeb138e4c01ba5b64bb22e5125d5c823c37f78",
    "scientific_commit": "fa60312f93ec859e42a57f383f5972b8df4cf807",
    "evidence_merge_commit": "df244eb414e68ad97c7bcce228b6129522e9a969",
    "contract_path": "benchmarks/q22-compositional-solomon-prospective-v1/contract.json",
    "contract_sha256": "f5fdd260ae7e7ef2fdab85bb4aaf0aaebb2716015b3e8026a677aaf8fa8d0ae4",
    "result_path": "benchmarks/q22-compositional-solomon-prospective-v1/evidence/result.json",
    "result_sha256": "c3b826685fc004fbc9b69a3a8ee539987522c1d3bf61e6e8d7c251147b394900",
    "evidence_manifest_sha256": "30db8f6f7669daaac8002a11b190e6e8a19d16ef40bdf988c1d0de92bac5ba72",
    "model_freeze_sha256": "928d912b56018d696ab2dda2497a5ca98c41f55eafe5ae1a63535d69c4fc9526",
    "evaluation_open_record_sha256": "a2c8e8a07b7bafa5376ff6fcffa5a0d3e4a645a87f18534e090593ae5f20a94e"
  },
  "firewall": {
    "models_frozen_before_evaluation": true,
    "training_had_evaluation_path": false,
    "prediction_received_blind_input_only": true,
    "retraining_allowed_after_open": false,
    "paid_compute_used": false
  },
  "scientific_run": {
    "seeds": [
      {
        "seed": 1,
        "cases": 1000,
        "operation_exact": 425,
        "operation_exact_rate_ppm": 425000,
        "per_class_exact_rate_ppm": {
          "quantity.add": 0,
          "quantity.multiply": 595000,
          "quantity.add-rational": 750000,
          "quantity.convert": 750000,
          "quantity.solve-linear": 30000
        },
        "model_sha256": "7e65210bc40be1728007a8b8ba9bf5c05983779e44dbaab8b240e533e51dec07"
      },
      {
        "seed": 2,
        "cases": 1000,
        "operation_exact": 430,
        "operation_exact_rate_ppm": 430000,
        "per_class_exact_rate_ppm": {
          "quantity.add": 165000,
          "quantity.multiply": 750000,
          "quantity.add-rational": 410000,
          "quantity.convert": 750000,
          "quantity.solve-linear": 75000
        },
        "model_sha256": "e73196d52784c2928e65326a50de035c7c9ffd19d79e96029bc0557acbbcc688"
      },
      {
        "seed": 3,
        "cases": 1000,
        "operation_exact": 533,
        "operation_exact_rate_ppm": 533000,
        "per_class_exact_rate_ppm": {
          "quantity.add": 250000,
          "quantity.multiply": 605000,
          "quantity.add-rational": 295000,
          "quantity.convert": 960000,
          "quantity.solve-linear": 555000
        },
        "model_sha256": "8c8f620ba4e8cca5b9e3503753523007d1f37b15eda14eb9b9e2bf0a3989da56"
      }
    ],
    "minimum_operation_exact_rate_ppm": 425000,
    "minimum_per_class_exact_rate_ppm": 0,
    "mean_operation_exact_rate_ppm": 462666,
    "minimum_margin_over_prefix_ppm": 225000,
    "all_seed_agreement_cases": 531,
    "all_seed_agreement_rate_ppm": 531000,
    "family_passed": false
  },
  "forecasts": {
    "ledger_recorded_settlement": false,
    "analytical_settlement": [
      {
        "forecast_id": "solomon.q22-compositional.forecast.mechanistic.v1",
        "probabilities": {"go": 0.72, "no_go": 0.27, "execution_failure": 0.01},
        "resolved_outcome": "no_go",
        "stake": 10,
        "brier_score": 1.0514
      },
      {
        "forecast_id": "solomon.q22-compositional.forecast.empirical.v1",
        "probabilities": {"go": 0.82, "no_go": 0.17, "execution_failure": 0.01},
        "resolved_outcome": "no_go",
        "stake": 10,
        "brier_score": 1.3614
      }
    ]
  },
  "interpretation": {
    "prefix_shortcut_removed": true,
    "held_out_template_transfer_passed": false,
    "some_operation_signal_learned": true,
    "arithmetic_answer_generation_tested": false,
    "general_language_quality_tested": false,
    "solomon_model_promotion_authorized": false,
    "next_boundary": "Close this fixed sparse class-head contract. Any successor must use a new frozen evaluation surface and change representation or objective rather than tune against the opened promotion set."
  }
}
