{
  "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": 0.02141550718806684
      },
      {
        "cfg": {
          "lr": 0.003
        },
        "mean": 0.0037718631356256083
      },
      {
        "cfg": {
          "lr": 0.01
        },
        "mean": 0.0034299326362088323
      }
    ],
    "full": {
      "mean": 0.002775895278318785,
      "std": 0.0017084856346187775,
      "per_seed": [
        0.0014615225372835994,
        0.004334237892180681,
        0.0012712624156847596,
        0.006652707699686289,
        0.0020110509358346462,
        0.0024756770581007004,
        0.0018914247630164027,
        0.0021092789247632027
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.0026888996071647853,
    "std": 0.0009078178093905949,
    "per_seed": [
      0.003874952672049403,
      0.001496415352448821,
      0.003612219588831067,
      0.0016158639919012785,
      0.0022634107153862715,
      0.0023024550173431635,
      0.002479603746905923,
      0.0038662757724523544
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": -8.699567115399987e-05,
    "idea_wins": 3,
    "n_pairs": 8,
    "per_seed_diffs": [
      0.002413430134765804,
      -0.00283782253973186,
      0.0023409571731463075,
      -0.00503684370778501,
      0.00025235977955162525,
      -0.0001732220407575369,
      0.0005881789838895202,
      0.0017569968476891518
    ],
    "p_value": 0.88165,
    "mde": 0.0021939824201293746,
    "mde_rel_pct": 79.03693043702121,
    "verdict": "no significant win",
    "system_worked": false
  },
  "idea_config": {
    "lr": 0.01
  },
  "idea_sweep": [
    {
      "cfg": {
        "lr": 0.001
      },
      "result": {
        "mean": 0.024548907764256,
        "std": 0.009325004444630938,
        "per_seed": [
          0.013507101684808731,
          0.021012168377637863,
          0.026467259973287582,
          0.027969608083367348,
          0.03438212722539902,
          0.0225788876414299,
          0.040141988545656204,
          0.010332120582461357
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.003
      },
      "result": {
        "mean": 0.007198455787147395,
        "std": 0.0027892140281787488,
        "per_seed": [
          0.0019068297697231174,
          0.010496499016880989,
          0.008480015210807323,
          0.006692741997539997,
          0.006822704337537289,
          0.005702317226678133,
          0.006068510934710503,
          0.011418027803301811
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.01
      },
      "result": {
        "mean": 0.0026888996071647853,
        "std": 0.0009078178093905949,
        "per_seed": [
          0.003874952672049403,
          0.001496415352448821,
          0.003612219588831067,
          0.0016158639919012785,
          0.0022634107153862715,
          0.0023024550173431635,
          0.002479603746905923,
          0.0038662757724523544
        ],
        "n": 8
      }
    }
  ],
  "protocol_note": "baseline sweep and idea sweep use identical learning-rate union; all task metrics are test MSE",
  "mechanism_signature": {
    "prediction": "integral coupling should yield small transverse hidden disagreement in trained heterogeneous copies",
    "observed_mean_pairwise_hidden_std": 0.02423221245408058,
    "per_seed": [
      0.02354380488395691,
      0.023498380556702614,
      0.024310877546668053,
      0.02557578682899475
    ],
    "threshold": 0.2,
    "confirmed": true
  }
}