{
  "bench_version": 1,
  "track": "incidence_actions",
  "model": "custom_incidence_head",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.006,
      "epochs": 12
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "epochs": 12
        },
        "mean": 0.7750000208616257
      },
      {
        "cfg": {
          "lr": 0.003,
          "epochs": 12
        },
        "mean": 0.6015625
      },
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12
        },
        "mean": 0.3687500059604645
      }
    ],
    "full": {
      "mean": 0.3562500048428774,
      "std": 0.06817944972613313,
      "per_seed": [
        0.32500001788139343,
        0.41875001788139343,
        0.3812499940395355,
        0.3499999940395355,
        0.3062500059604645,
        0.22500000894069672,
        0.4625000059604645,
        0.3812499940395355
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.12421874981373549,
    "std": 0.02854504035137365,
    "per_seed": [
      0.09375,
      0.15000000596046448,
      0.14374999701976776,
      0.11874999850988388,
      0.15625,
      0.06875000149011612,
      0.14374999701976776,
      0.11874999850988388
    ],
    "n": 8,
    "best_cfg": {
      "lr": 0.006,
      "epochs": 12,
      "alpha": 0.0
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 0.0
        },
        "mean": 0.12656250037252903
      },
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 0.5
        },
        "mean": 0.27812499552965164
      },
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 1.0
        },
        "mean": 0.690625011920929
      }
    ]
  },
  "comparison": {
    "delta_mean": -0.2320312550291419,
    "idea_wins": 8,
    "n_pairs": 8,
    "per_seed_diffs": [
      -0.23125001788139343,
      -0.26875001192092896,
      -0.23749999701976776,
      -0.23124999552965164,
      -0.15000000596046448,
      -0.1562500074505806,
      -0.3187500089406967,
      -0.26249999552965164
    ],
    "p_value": 0.0081,
    "mde": 0.04719409486255682,
    "mde_rel_pct": 13.247465044490758,
    "verdict": "idea better (significant)",
    "system_worked": true
  },
  "mechanism_signature": {
    "custom_track": {
      "name": "incidence_actions",
      "file": "incidence_actions.py",
      "domain": "structured_action_classification"
    },
    "baseline_sweep_union_lrs": [
      0.001,
      0.003,
      0.006
    ],
    "idea_sweep": [
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 0.0
        },
        "mean": 0.12656250037252903
      },
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 0.5
        },
        "mean": 0.27812499552965164
      },
      {
        "cfg": {
          "lr": 0.006,
          "epochs": 12,
          "alpha": 1.0
        },
        "mean": 0.690625011920929
      }
    ],
    "mechanism_signature": {
      "prediction": "trained incidence logits equal A times trained atomic scores (alpha-normalized)",
      "n_observed_logits": 1636,
      "predicted_max_abs_error": 0.0,
      "observed_mean_abs_error": 0.0,
      "confirmed": true
    }
  }
}