{
  "bench_version": 1,
  "track": "tabular",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "wd": 0.0
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "wd": 0.0
        },
        "mean": 14.000884532928467
      },
      {
        "cfg": {
          "lr": 0.001,
          "wd": 0.0001
        },
        "mean": 13.928702116012573
      },
      {
        "cfg": {
          "lr": 0.001,
          "wd": 0.001
        },
        "mean": 13.853612661361694
      },
      {
        "cfg": {
          "lr": 0.003,
          "wd": 0.0
        },
        "mean": 7.579635977745056
      },
      {
        "cfg": {
          "lr": 0.003,
          "wd": 0.0001
        },
        "mean": 7.295058488845825
      },
      {
        "cfg": {
          "lr": 0.003,
          "wd": 0.001
        },
        "mean": 6.986483931541443
      },
      {
        "cfg": {
          "lr": 0.01,
          "wd": 0.0
        },
        "mean": 6.920138120651245
      },
      {
        "cfg": {
          "lr": 0.01,
          "wd": 0.0001
        },
        "mean": 6.949732184410095
      },
      {
        "cfg": {
          "lr": 0.01,
          "wd": 0.001
        },
        "mean": 6.989647626876831
      }
    ],
    "full": {
      "mean": 6.9205562472343445,
      "std": 0.5716103719461259,
      "per_seed": [
        7.096155643463135,
        6.974558353424072,
        6.666646480560303,
        6.943192005157471,
        5.7007622718811035,
        7.199983596801758,
        6.8888444900512695,
        7.8943071365356445
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 7.772280991077423,
    "std": 0.6641172020856825,
    "per_seed": [
      7.700952529907227,
      7.535534381866455,
      8.856191635131836,
      6.979861736297607,
      6.709100246429443,
      7.994025707244873,
      8.44301700592041,
      7.959564685821533
    ],
    "n": 8,
    "chosen_cfg": {
      "lr": 0.003,
      "lam": 1e-05,
      "eta_up": 1.5
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "lam": 1e-05,
          "eta_up": 1.35
        },
        "mean": 8.174783706665039,
        "per_seed": [
          8.012105941772461,
          8.44368839263916,
          9.441221237182617,
          7.554927825927734,
          7.6160888671875,
          8.013986587524414,
          8.08622932434082,
          8.230021476745605
        ]
      },
      {
        "cfg": {
          "lr": 0.003,
          "lam": 1e-05,
          "eta_up": 1.5
        },
        "mean": 7.772280991077423,
        "per_seed": [
          7.700952529907227,
          7.535534381866455,
          8.856191635131836,
          6.979861736297607,
          6.709100246429443,
          7.994025707244873,
          8.44301700592041,
          7.959564685821533
        ]
      },
      {
        "cfg": {
          "lr": 0.01,
          "lam": 1e-05,
          "eta_up": 1.5
        },
        "mean": 7.991110026836395,
        "per_seed": [
          8.947175025939941,
          7.837967872619629,
          7.783258438110352,
          7.059815406799316,
          7.544052600860596,
          7.272561073303223,
          7.9431962966918945,
          9.540853500366211
        ]
      }
    ]
  },
  "comparison": {
    "delta_mean": 0.8517247438430786,
    "idea_wins": 0,
    "n_pairs": 8,
    "per_seed_diffs": [
      0.6047968864440918,
      0.5609760284423828,
      2.189545154571533,
      0.03666973114013672,
      1.0083379745483398,
      0.7940421104431152,
      1.5541725158691406,
      0.06525754928588867
    ],
    "p_value": 0.0081,
    "mde": 0.611136711877834,
    "mde_rel_pct": 8.830745536127418,
    "verdict": "idea worse (significant)",
    "system_worked": false
  },
  "mechanism_signature": {
    "prediction": "adaptive eta should remain below local secant stability boundary 2/kappa",
    "predicted_eta_c_from_trained_secants": 6.806255525998862e-09,
    "observed_max_final_eta": 0.00043937891010204985,
    "observed_median_eta": 0.0004584609905860991,
    "accepted_mean": 124.25,
    "rejected_mean": 25.75,
    "confirmed": false
  },
  "notes": "Matched Friedman#1 regression and shared mlp_tiny; optimizer is the only intervention. Baseline AdamW sweep covers lr and weight decay."
}