{
  "benchmark": "planbench-mini",
  "seed": 42,
  "episodes": 120,
  "tapo": {
    "joint_success": 1.0,
    "average_cost": 1.05,
    "transition_targets": 360,
    "transition_accuracy": 1.0
  },
  "grsd": {
    "joint_success": 1.0,
    "average_cost": 1.35,
    "reflective_groups": 120,
    "success_failure_contrasts": 120,
    "privileged_guidance_updates": 360
  },
  "fidelity": "deterministic mini-suite mechanism reproduction",
  "schema_version": 2,
  "manifest_ref": "experiments:tapo-grsd-planbench-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/tapo-grsd-planbench-seed42.json",
    "historical_migration": "historical-metrics-v2-2026-08-09",
    "original_code_commit": "not recorded",
    "dataset_fingerprint": "not recorded in historical artifact"
  }
}