{
  "bench_version": 1,
  "track": "dynamics",
  "model": "rnn_small",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001
        },
        "mean": 0.0008724518556846306
      },
      {
        "cfg": {
          "lr": 0.003
        },
        "mean": 0.00035974886850453913
      },
      {
        "cfg": {
          "lr": 0.01
        },
        "mean": 4.963686433256953e-05
      }
    ],
    "full": {
      "mean": 4.311938073442434e-05,
      "std": 2.1245040862825782e-05,
      "per_seed": [
        9.722933464217931e-05,
        3.783951615332626e-05,
        3.003067286044825e-05,
        3.3447933674324304e-05,
        4.86072531202808e-05,
        2.836398380168248e-05,
        3.394361920072697e-05,
        3.549273242242634e-05
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 4.2328419567638775e-05,
    "std": 2.1386848202922675e-05,
    "per_seed": [
      9.703257819637656e-05,
      3.75409254047554e-05,
      2.9079366868245415e-05,
      3.262508107582107e-05,
      4.653289943234995e-05,
      2.7314470571582206e-05,
      3.378422115929425e-05,
      3.471781383268535e-05
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": -7.909611667855643e-07,
    "idea_wins": 8,
    "n_pairs": 8,
    "per_seed_diffs": [
      -1.967564458027482e-07,
      -2.9859074857085943e-07,
      -9.513059922028333e-07,
      -8.22852598503232e-07,
      -2.074353687930852e-06,
      -1.049513230100274e-06,
      -1.593980414327234e-07,
      -7.749185897409916e-07
    ],
    "p_value": 0.0081,
    "mde": 5.226854739698207e-07,
    "mde_rel_pct": 1.2121822370063282,
    "verdict": "idea better (significant)",
    "system_worked": true
  },
  "mechanism_signature": {
    "prediction": "orthogonal transitions preserve hidden norms and patch loss reduces disagreement",
    "predicted_norm_error": 0.0,
    "observed_max_norm_error": 2.384185791015625e-07,
    "predicted_disagreement_change": "negative",
    "observed_baseline_disagreement": 0.002737239032285288,
    "observed_idea_disagreement": 0.002110361514496617,
    "confirmed": true,
    "math_sanity": {
      "max_norm_error": 4.76837158203125e-07,
      "predicted": "orthogonal gauge preserves hidden norm"
    },
    "trained_model_behavior": {
      "baseline": [
        {
          "metric": 9.722933464217931e-05,
          "disagreement": 0.0028517728205770254,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 3.783951615332626e-05,
          "disagreement": 0.0026706315111368895,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 3.003067286044825e-05,
          "disagreement": 0.0027825413271784782,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 3.3447933674324304e-05,
          "disagreement": 0.002656190888956189,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 4.86072531202808e-05,
          "disagreement": 0.002491437830030918,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 2.836398380168248e-05,
          "disagreement": 0.003299158299341798,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 3.394361920072697e-05,
          "disagreement": 0.0026531326584517956,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        },
        {
          "metric": 3.549273242242634e-05,
          "disagreement": 0.00249304692260921,
          "norm_error": 0.0,
          "angle_mean_abs": 0.0
        }
      ],
      "idea": [
        {
          "metric": 9.703257819637656e-05,
          "disagreement": 0.0022101164795458317,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.05565014109015465
        },
        {
          "metric": 3.75409254047554e-05,
          "disagreement": 0.0020225902553647757,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.06866727769374847
        },
        {
          "metric": 2.9079366868245415e-05,
          "disagreement": 0.002208986319601536,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.05578428506851196
        },
        {
          "metric": 3.262508107582107e-05,
          "disagreement": 0.0020868959836661816,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.048349007964134216
        },
        {
          "metric": 4.653289943234995e-05,
          "disagreement": 0.001802699756808579,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.053223490715026855
        },
        {
          "metric": 2.7314470571582206e-05,
          "disagreement": 0.0024720188230276108,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.056646957993507385
        },
        {
          "metric": 3.378422115929425e-05,
          "disagreement": 0.002076399978250265,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.05470898002386093
        },
        {
          "metric": 3.471781383268535e-05,
          "disagreement": 0.0020031845197081566,
          "norm_error": 2.384185791015625e-07,
          "angle_mean_abs": 0.054334621876478195
        }
      ]
    },
    "idea_sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "lambda_patch": 0.01
        },
        "result": {
          "mean": 0.0009701697272248566,
          "std": 0.0002551454195727328,
          "per_seed": [
            0.0006990438560023904,
            0.0009627738618291914,
            0.0007161990506574512,
            0.0011117245303466916,
            0.0007498612976633012,
            0.0008323467336595058,
            0.001259242300875485,
            0.0014301661867648363
          ],
          "n": 8
        }
      },
      {
        "cfg": {
          "lr": 0.003,
          "lambda_patch": 0.01
        },
        "result": {
          "mean": 0.0003214057542209048,
          "std": 5.179809947448413e-05,
          "per_seed": [
            0.0003400232526473701,
            0.0004083190578967333,
            0.00031012375256977975,
            0.00038058325299061835,
            0.0003361486305948347,
            0.00027818765374831855,
            0.0002542929141782224,
            0.0002635675191413611
          ],
          "n": 8
        }
      },
      {
        "cfg": {
          "lr": 0.01,
          "lambda_patch": 0.01
        },
        "result": {
          "mean": 4.2328419567638775e-05,
          "std": 2.1386848202922675e-05,
          "per_seed": [
            9.703257819637656e-05,
            3.75409254047554e-05,
            2.9079366868245415e-05,
            3.262508107582107e-05,
            4.653289943234995e-05,
            2.7314470571582206e-05,
            3.378422115929425e-05,
            3.471781383268535e-05
          ],
          "n": 8
        }
      }
    ],
    "track_rationale": "Dynamics is the built-in stability/control track; local temporal patches are neighboring state windows."
  }
}