{
  "bench_version": 1,
  "track": "dynamics",
  "model": "rnn_small",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.006
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001
        },
        "mean": 0.0002686133266252
      },
      {
        "cfg": {
          "lr": 0.003
        },
        "mean": 2.2429956516134553e-05
      },
      {
        "cfg": {
          "lr": 0.006
        },
        "mean": 1.2719097412627889e-05
      }
    ],
    "full": {
      "mean": 1.3684802866009704e-05,
      "std": 2.4297331089709535e-06,
      "per_seed": [
        1.0965275578200817e-05,
        1.3063688129477669e-05,
        1.363205592497252e-05,
        1.321537001786055e-05,
        1.694267484708689e-05,
        1.790331407391932e-05,
        1.3341849808057304e-05,
        1.0414194548502564e-05
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.0005339367999113165,
    "std": 0.0003344099304737304,
    "per_seed": [
      0.00038349704118445516,
      0.0002837758220266551,
      0.00047679623821750283,
      0.000840833701658994,
      0.00030266467365436256,
      0.00031309283804148436,
      0.0012979843886569142,
      0.0003728496958501637
    ],
    "n": 8,
    "chosen_cfg": {
      "lr": 0.006,
      "alpha": 0.75
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "alpha": 0.25
        },
        "mean": 0.5652050226926804,
        "per_seed": [
          0.5019805431365967,
          0.6064469218254089,
          0.5146586894989014,
          0.6377339363098145
        ]
      },
      {
        "cfg": {
          "lr": 0.003,
          "alpha": 0.5
        },
        "mean": 0.05728436540812254,
        "per_seed": [
          0.0478343702852726,
          0.05792839825153351,
          0.03978803753852844,
          0.08358665555715561
        ]
      },
      {
        "cfg": {
          "lr": 0.006,
          "alpha": 0.75
        },
        "mean": 0.0004962257007719018,
        "per_seed": [
          0.00038349704118445516,
          0.0002837758220266551,
          0.00047679623821750283,
          0.000840833701658994
        ]
      }
    ]
  },
  "comparison": {
    "delta_mean": 0.0005202519970453068,
    "idea_wins": 0,
    "n_pairs": 8,
    "per_seed_diffs": [
      0.00037253176560625434,
      0.0002707121338971774,
      0.0004631641822925303,
      0.0008276183316411334,
      0.00028572199880727567,
      0.00029518952396756504,
      0.001284642538848857,
      0.00036243550130166113
    ],
    "p_value": 0.0081,
    "mde": 0.00029926510858487807,
    "mde_rel_pct": 2186.8426715023597,
    "verdict": "idea worse (significant)",
    "system_worked": false
  },
  "mechanism_signature": {
    "prediction": "relaxation damps stochastic block-response updates approximately linearly in alpha",
    "predicted_alpha": 0.75,
    "observed_update_rms_mean": 0.0026733302263543613,
    "observed_update_std_mean": 0.0019501892436788798,
    "predicted_vs_observed_relation": "measured from trained rnn_small updates; no analytic/toy data used",
    "confirmed": false,
    "confirmation_note": "Not quantitatively confirmed: only one alpha was evaluated in the final trained-model signature, so linear alpha scaling was not tested."
  },
  "custom_track": null
}