{
  "carry_vs_bag": {
    "available": true,
    "bag_primary_accuracy": 0.296875,
    "carry_answer_interface_valid": true,
    "carry_full_top_is_answer_rate": 1.0,
    "carry_full_top_is_answer_rate_per_seed": {
      "7401": 1.0
    },
    "carry_minus_bag": 0.04296875,
    "ci95": [
      -0.0078125,
      0.09375
    ],
    "complete_primary_depths": 8,
    "expected_model_seeds": [
      7401
    ],
    "expected_unique_tasks_per_primary_depth": 32,
    "model_seed_count": 1,
    "model_seeds": [
      7401
    ],
    "per_depth": {
      "10": {
        "ci95": [
          -0.125,
          0.125
        ],
        "difference": 0.0,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0
        },
        "unique_tasks": 32
      },
      "11": {
        "ci95": [
          -0.0625,
          0.1875
        ],
        "difference": 0.0625,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0625
        },
        "unique_tasks": 32
      },
      "12": {
        "ci95": [
          -0.25,
          0.0625
        ],
        "difference": -0.09375,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": -0.09375
        },
        "unique_tasks": 32
      },
      "5": {
        "ci95": [
          -0.09375,
          0.21875
        ],
        "difference": 0.0625,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0625
        },
        "unique_tasks": 32
      },
      "6": {
        "ci95": [
          0.09375,
          0.375
        ],
        "difference": 0.21875,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.21875
        },
        "unique_tasks": 32
      },
      "7": {
        "ci95": [
          -0.03125,
          0.21875
        ],
        "difference": 0.09375,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.09375
        },
        "unique_tasks": 32
      },
      "8": {
        "ci95": [
          -0.0625,
          0.1875
        ],
        "difference": 0.0625,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0625
        },
        "unique_tasks": 32
      },
      "9": {
        "ci95": [
          -0.1875,
          0.0625
        ],
        "difference": -0.0625,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": -0.0625
        },
        "unique_tasks": 32
      }
    },
    "per_query_kind": {
      "checksum": {
        "ci95": [
          -0.015625,
          0.140625
        ],
        "difference": 0.0625,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0625
        },
        "unique_tasks": 128
      },
      "node": {
        "ci95": [
          -0.03125,
          0.0859375
        ],
        "difference": 0.0234375,
        "model_seed_count": 1,
        "model_seeds": [
          7401
        ],
        "per_seed": {
          "7401": 0.0234375
        },
        "unique_tasks": 128
      }
    },
    "per_seed": {
      "7401": 0.04296875
    },
    "positive_depths": 5,
    "query_kinds_complete": true,
    "query_kinds_positive": true,
    "required_primary_depths": 8,
    "seed_receipts": [
      {
        "checkpoint_phase": "pilot",
        "initialization_equal": true,
        "post_checkpoint_k1_parity": true,
        "seed": 7401,
        "training_compute_equal": true,
        "training_order_equal": true,
        "unique_primary_tasks": 256
      }
    ],
    "seed_set_complete": true,
    "unique_tasks": 256
  },
  "config_sha256": "70e4a2d6df7acb0c5a21c7c945c66499a0ede8e98321c7b56da1c080c819744b",
  "counterfactual_swaps": {
    "available": true,
    "complete": true,
    "expected_model_seeds": [
      7401
    ],
    "model_seeds": [
      7401
    ],
    "passes": false,
    "seeds": [
      {
        "baseline_donor_follow_rate": 0.2109375,
        "bootstrap_unit": "counterfactual_pair_mean_over_two_directions",
        "ci95": [
          -0.0234375,
          0.0390625
        ],
        "directions": 128,
        "donor_follow_gain": 0.0078125,
        "donor_follow_minus_recipient_preserve": -0.0546875,
        "post_swap_donor_follow_rate": 0.21875,
        "post_swap_recipient_preserve_rate": 0.2734375,
        "train_seed": 7401,
        "unique_pairs": 64
      }
    ]
  },
  "data_manifest_sha256": "2cf9a4d008d0990873928424170ab5daf0a53473a9f97e542d8739ac6de92879",
  "deployment_comparison": {
    "available": false,
    "reason": "deployment comparator is prohibited in the pilot phase"
  },
  "evaluation_bundles": 2,
  "experiment_id": "qwen35_4b_state_carry_vs_state_bag",
  "joint_holdout_carry_vs_bag": {
    "available": true,
    "carry_minus_bag": 0.05078125,
    "ci95": [
      0.0078125,
      0.09765625
    ],
    "complete_depths": 8,
    "expected_unique_tasks_per_depth": 32,
    "model_seed_count": 1,
    "model_seeds": [
      7401
    ],
    "non_reversal_all_seeds": true,
    "passes": true,
    "per_seed": {
      "7401": 0.05078125
    },
    "required_depths": 8,
    "seed_set_complete": true,
    "unique_tasks": 256
  },
  "phase": "pilot",
  "pilot_gate": {
    "checks": {
      "answer_interface_valid": true,
      "complete_registered_cells": true,
      "gate_reachable": true,
      "joint_holdout_diagnostic_complete": true,
      "joint_state_sufficient": false,
      "k4_diagnostic_complete": true,
      "positive_carry_minus_bag": true,
      "query_kinds_positive": true,
      "swap_diagnostic_complete": true
    },
    "complete": true,
    "expected_model_seed": 7401,
    "failure_reason": null,
    "promote": false,
    "reachability": {
      "baseline": 0.296875,
      "reachable": true,
      "required_gain": 0.05,
      "required_score": 0.346875,
      "upper_bound": 1.0
    },
    "status": "PILOT_MECHANISM_MISS"
  },
  "receipt_identity_sha256": "c9fec5b584bef5aa3de40844a8552395f3c3f0f95f79285700a9e4e375aed0ef",
  "requirements_training_lock_sha256": "05546fe977583116d6169ea0dfa7b27e1184dd4a2b61d556dfb3f889d5b2b7b1",
  "sample_more": {
    "available": false,
    "reason": "deployment comparator is prohibited in the pilot phase"
  },
  "schema_version": 2,
  "source_contract_sha256": "ef2dd25107cb306490e30dba8ac3035c8c69c76173cec62390557cf1add7a28d",
  "state_sufficiency": {
    "available": true,
    "joint_step_accuracy": 0.0045947759645059705,
    "model_seeds": [
      7401
    ],
    "node_step_accuracy": 0.06419115958851762,
    "passes": false,
    "unique_tasks": 256
  },
  "trained_checkpoint_edge_cut": {
    "available": false,
    "model_seeds": []
  },
  "unseen_k_scaling": {
    "available": true,
    "ci95": [
      -0.03515625,
      0.05859375
    ],
    "complete": true,
    "expected_unique_tasks": 256,
    "gain": 0.01171875,
    "model_seed_count": 1,
    "model_seeds": [
      7401
    ],
    "per_seed": {
      "7401": 0.01171875
    },
    "unique_tasks": 256
  },
  "verdict": "PILOT_MECHANISM_MISS",
  "warning": "Only DEPLOYABLE_DEPTH_BREAKTHROUGH licenses a deployment claim; mechanistic labels remain valuable but narrower."
}