{
  "bench_version": 1,
  "track": "tabular",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "weight_decay": 0.0
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "weight_decay": 0.0
        },
        "mean": 17.79831099510193
      },
      {
        "cfg": {
          "lr": 0.003,
          "weight_decay": 0.0
        },
        "mean": 13.153547525405884
      },
      {
        "cfg": {
          "lr": 0.01,
          "weight_decay": 0.0
        },
        "mean": 7.754182696342468
      }
    ],
    "full": {
      "mean": 7.931941092014313,
      "std": 0.8101161551333251,
      "per_seed": [
        7.2100510597229,
        8.814193725585938,
        7.362153053283691,
        7.630332946777344,
        6.520580291748047,
        8.472823143005371,
        8.71557903289795,
        8.729815483093262
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 8.868400931358337,
    "std": 0.7219192817226784,
    "per_seed": [
      9.187010765075684,
      8.92486572265625,
      8.406270980834961,
      8.585136413574219,
      7.358145713806152,
      9.900330543518066,
      9.11246395111084,
      9.472983360290527
    ],
    "n": 8,
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "weight_decay": 0.0
        },
        "mean": 44.67299127578735
      },
      {
        "cfg": {
          "lr": 0.003,
          "weight_decay": 0.0
        },
        "mean": 12.883334636688232
      },
      {
        "cfg": {
          "lr": 0.01,
          "weight_decay": 0.0
        },
        "mean": 8.775820970535278
      }
    ]
  },
  "comparison": {
    "delta_mean": 0.9364598393440247,
    "idea_wins": 0,
    "n_pairs": 8,
    "per_seed_diffs": [
      1.9769597053527832,
      0.1106719970703125,
      1.0441179275512695,
      0.954803466796875,
      0.8375654220581055,
      1.4275074005126953,
      0.3968849182128906,
      0.7431678771972656
    ],
    "p_value": 0.0081,
    "mde": 0.4850827510411753,
    "mde_rel_pct": 6.115561694344213,
    "verdict": "idea worse (significant)",
    "system_worked": false
  },
  "mechanism_signature": {
    "prediction": "idea starts in recovered spectral subspace, so projection energy is higher",
    "predicted_vs_observed": {
      "observed_energy_ratio": 1.7969543217617925,
      "threshold": 1.2
    },
    "confirmed": true,
    "trained_model_measurements": [
      {
        "seed": 0,
        "spectral_gap": 2.3793046737761143,
        "baseline_projection_energy": 0.153251051902771,
        "idea_projection_energy": 0.28242650628089905,
        "idea_metric": 9.187010765075684
      },
      {
        "seed": 1,
        "spectral_gap": 2.345487753974355,
        "baseline_projection_energy": 0.14754755795001984,
        "idea_projection_energy": 0.2536274790763855,
        "idea_metric": 8.92486572265625
      },
      {
        "seed": 2,
        "spectral_gap": 2.5917119317759125,
        "baseline_projection_energy": 0.1287820041179657,
        "idea_projection_energy": 0.2661108076572418,
        "idea_metric": 8.406270980834961
      },
      {
        "seed": 3,
        "spectral_gap": 2.0604149370605582,
        "baseline_projection_energy": 0.15506194531917572,
        "idea_projection_energy": 0.28266391158103943,
        "idea_metric": 8.585136413574219
      },
      {
        "seed": 4,
        "spectral_gap": 2.544887882094244,
        "baseline_projection_energy": 0.17158180475234985,
        "idea_projection_energy": 0.24406076967716217,
        "idea_metric": 7.358145713806152
      },
      {
        "seed": 5,
        "spectral_gap": 2.366815131166324,
        "baseline_projection_energy": 0.1343323141336441,
        "idea_projection_energy": 0.26110315322875977,
        "idea_metric": 9.900330543518066
      },
      {
        "seed": 6,
        "spectral_gap": 2.2032378706059492,
        "baseline_projection_energy": 0.16026736795902252,
        "idea_projection_energy": 0.2490236610174179,
        "idea_metric": 9.11246395111084
      },
      {
        "seed": 7,
        "spectral_gap": 2.286019756494266,
        "baseline_projection_energy": 0.13112203776836395,
        "idea_projection_energy": 0.2628439664840698,
        "idea_metric": 9.472983360290527
      }
    ]
  },
  "protocol": {
    "paired_seeds": [
      0,
      1,
      2,
      3,
      4,
      5,
      6,
      7
    ],
    "n_train": 400,
    "n_test": 400,
    "epochs": 20,
    "batch": 128,
    "common_grid": [
      {
        "lr": 0.001,
        "weight_decay": 0.0
      },
      {
        "lr": 0.003,
        "weight_decay": 0.0
      },
      {
        "lr": 0.01,
        "weight_decay": 0.0
      }
    ],
    "track_rationale": "Initialization intervention uses the designated tabular MLP track."
  }
}