{
  "bench_version": 1,
  "track": "strain_congruence",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "weight_decay": 0.0001
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "weight_decay": 0.0
        },
        "mean": 0.07749666459858418
      },
      {
        "cfg": {
          "lr": 0.003,
          "weight_decay": 0.0
        },
        "mean": 0.05831924267113209
      },
      {
        "cfg": {
          "lr": 0.01,
          "weight_decay": 0.0001
        },
        "mean": 0.033658347092568874
      }
    ],
    "full": {
      "mean": 0.029577972134575248,
      "std": 0.01503937321913688,
      "per_seed": [
        0.021902937442064285,
        0.026272417977452278,
        0.06678790599107742,
        0.01967012695968151,
        0.02644897624850273,
        0.02151375077664852,
        0.03620384261012077,
        0.01782381907105446
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.00675772619433701,
    "std": 0.0009987191335145192,
    "per_seed": [
      0.005988644901663065,
      0.0065969908609986305,
      0.005767101421952248,
      0.008887243457138538,
      0.007409937214106321,
      0.006973727606236935,
      0.006849044468253851,
      0.005589119624346495
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": -0.022820245940238237,
    "idea_wins": 8,
    "n_pairs": 8,
    "per_seed_diffs": [
      -0.01591429254040122,
      -0.019675427116453648,
      -0.061020804569125175,
      -0.010782883502542973,
      -0.01903903903439641,
      -0.014540023170411587,
      -0.029354798141866922,
      -0.012234699446707964
    ],
    "p_value": 0.0081,
    "mde": 0.01377344810526238,
    "mde_rel_pct": 46.566573403326295,
    "verdict": "idea better (significant)",
    "system_worked": true
  },
  "idea_sweep": [
    {
      "cfg": {
        "lr": 0.001,
        "weight_decay": 0.0
      },
      "result": {
        "mean": 0.009227964736055583,
        "std": 0.0011620368930055716,
        "per_seed": [
          0.008295468986034393,
          0.009248356334865093,
          0.00782246794551611,
          0.010978559963405132,
          0.010134941898286343,
          0.010268867015838623,
          0.009545788168907166,
          0.007529267575591803
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.003,
        "weight_decay": 0.0
      },
      "result": {
        "mean": 0.007636490918230265,
        "std": 0.0011627000336291112,
        "per_seed": [
          0.006451956927776337,
          0.009120338596403599,
          0.006323990877717733,
          0.009338542819023132,
          0.007833528332412243,
          0.008401229046285152,
          0.007360921707004309,
          0.006261419039219618
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.01,
        "weight_decay": 0.0001
      },
      "result": {
        "mean": 0.00675772619433701,
        "std": 0.0009987191335145192,
        "per_seed": [
          0.005988644901663065,
          0.0065969908609986305,
          0.005767101421952248,
          0.008887243457138538,
          0.007409937214106321,
          0.006973727606236935,
          0.006849044468253851,
          0.005589119624346495
        ],
        "n": 8
      }
    }
  ],
  "custom_track": {
    "name": "strain_congruence",
    "file": "custom_strain_track.py",
    "domain": "dynamics"
  },
  "mechanism_signature": {
    "prediction": "icosahedral reconstruction should reduce learned prediction error under ill-conditioned SL(3) deformations",
    "condition_bins": {
      "low": {
        "baseline_mse": 0.010883176291827112,
        "idea_mse": 0.008369530783966184
      },
      "mid": {
        "baseline_mse": 0.008730645466130227,
        "idea_mse": 0.0025046369701158255
      },
      "high": {
        "baseline_mse": 0.05419394630007446,
        "idea_mse": 0.01061116874916479
      }
    },
    "confirmed": true
  }
}