{
  "batch": "2026-08-08-p0-p1-closed-audit",
  "seed": 42,
  "post_training": {
    "distilled-rl": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.609375,
        "mean_reward": 0.6957341455073699,
        "entropy": 1.5622097613615191,
        "kl_from_reference": 0.22954970786053555
      },
      "relative_accuracy": 1.6,
      "diagnostics": {
        "reverse_ratio_clip_rate": 1.0,
        "negative_sample_resets": 5.0,
        "sequence_geometric_normalizer": 1.5874010519693902,
        "unconditional_teacher_matching": 0.0,
        "loss": 1.5223757919820429,
        "policy_entropy": 1.5749032225450217
      }
    },
    "u-opsd": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.09375,
        "mean_reward": 0.2513358801149202,
        "entropy": 1.3135866965242053,
        "kl_from_reference": 0.4781727726978496
      },
      "relative_accuracy": -0.6,
      "diagnostics": {
        "self_consistency_votes": 8.0,
        "pseudo_solution_confidence": 0.375,
        "external_supervision": 0.0,
        "repair_target_is_gold": 0.0,
        "loss": 1.0697468577895208,
        "policy_entropy": 1.2763636098947497
      }
    },
    "rp-opsd": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.59375,
        "mean_reward": 0.6766101787372542,
        "entropy": 1.668932321536027,
        "kl_from_reference": 0.12282714768602855
      },
      "relative_accuracy": 1.5333333333333334,
      "diagnostics": {
        "reasoning_pivot_mass": 0.7492843102945891,
        "privileged_positions": 3.0,
        "reference_anchor": 0.45,
        "loss": 1.439587945410726,
        "policy_entropy": 1.6551894534897618
      }
    },
    "pcsd": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.5625,
        "mean_reward": 0.6682015450116028,
        "entropy": 1.7492900489776715,
        "kl_from_reference": 0.04246942024438405
      },
      "relative_accuracy": 1.4,
      "diagnostics": {
        "persistent_gate_mean": 0.2586930518527548,
        "adaptive_window": 4.0,
        "trend_attenuated_positions": 2.0,
        "loss": 1.7163194025488961,
        "policy_entropy": 1.747686238904023
      }
    },
    "adrs": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.625,
        "mean_reward": 0.7120526941289935,
        "entropy": 1.6014089735062695,
        "kl_from_reference": 0.19035049571578586
      },
      "relative_accuracy": 1.6666666666666667,
      "diagnostics": {
        "teacher_value_advantage_gate": 0.7615812657124454,
        "within_step_score_mean": 4.625929269271486e-17,
        "return_teacher_association": 0.7615812657124454,
        "inference_time_skill": 0.0,
        "loss": 0.4205092193520913,
        "policy_entropy": 1.6319840744521554
      }
    },
    "mopd": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.34375,
        "mean_reward": 0.5027255717244984,
        "entropy": 1.747822304024646,
        "kl_from_reference": 0.04393716519740872
      },
      "relative_accuracy": 0.4666666666666667,
      "diagnostics": {
        "domain_teachers": 4.0,
        "student_rollout_support": 0.6666666666666666,
        "teacher_merge_parameters": 0.0,
        "largest_domain_weight": 0.25,
        "loss": 1.7270830842942804,
        "policy_entropy": 1.7632612481603198
      }
    },
    "opd-lm": {
      "baseline": {
        "accuracy": 0.234375,
        "mean_reward": 0.38409619661640243,
        "entropy": 1.7917594692220566,
        "kl_from_reference": 0.0
      },
      "final": {
        "accuracy": 0.65625,
        "mean_reward": 0.7323539145223933,
        "entropy": 1.7528317478975353,
        "kl_from_reference": 0.038927721324519896
      },
      "relative_accuracy": 1.8,
      "diagnostics": {
        "bidirectional_teacher": 1.0,
        "autoregressive_anchor": 0.7,
        "diffusion_denoising_views": 2.0,
        "teacher_trainable": 0.0,
        "loss": 1.7187243454533376,
        "policy_entropy": 1.770725022339059
      }
    }
  },
  "agent": {
    "agent-opsd": {
      "metrics": {
        "answer_accuracy": 1.0,
        "plan_success": 1.0,
        "joint_success": 1.0,
        "average_cost": 0.9000000000000006
      },
      "diagnostics": {
        "episodes": 120,
        "memory_size": 24,
        "actions": 360,
        "policy_updates": 120,
        "dense_credit_updates": 360,
        "turn_credit_updates": 360,
        "recursive_belief_updates": 360,
        "pivotal_turns": 200
      }
    },
    "ocsd": {
      "metrics": {
        "answer_accuracy": 1.0,
        "plan_success": 1.0,
        "joint_success": 1.0,
        "average_cost": 1.0200000000000005
      },
      "diagnostics": {
        "episodes": 120,
        "memory_size": 24,
        "actions": 360,
        "policy_updates": 120,
        "dense_credit_updates": 360,
        "turn_credit_updates": 360,
        "observation_calibrations": 360,
        "scaffold_ablations": 360
      }
    },
    "vermem": {
      "metrics": {
        "answer_accuracy": 1.0,
        "plan_success": 1.0,
        "joint_success": 1.0,
        "average_cost": 0.6099999999999995
      },
      "diagnostics": {
        "episodes": 120,
        "memory_size": 24,
        "reused_plans": 108,
        "actions": 360,
        "memories_retrieved": 108,
        "local_verifier_calls": 240,
        "global_verifier_calls": 120,
        "memory_operations": 240
      }
    },
    "coevo-mem": {
      "metrics": {
        "answer_accuracy": 1.0,
        "plan_success": 1.0,
        "joint_success": 1.0,
        "average_cost": 0.5
      },
      "diagnostics": {
        "episodes": 120,
        "memory_size": 24,
        "reused_plans": 54,
        "actions": 360,
        "policy_updates": 120,
        "coevolution_alternations": 120,
        "router_updates": 60,
        "memory_bank_updates": 60
      }
    }
  },
  "schema_version": 2,
  "manifest_ref": "experiments:p0-p1-closed-audit-20260808-seed42",
  "evaluation_protocol": {
    "tier": "l1_mechanism",
    "seeds": [
      42
    ],
    "formal_comparison": false,
    "claim_policy": "single/few-seed smoke result; do not claim a stable improvement"
  },
  "provenance": {
    "artifact_path": "docs/experiments/p0-p1-closed-audit-20260808-seed42.json",
    "historical_migration": "historical-metrics-v2-2026-08-09",
    "original_code_commit": "not recorded",
    "dataset_fingerprint": "not recorded in historical artifact"
  }
}