{
  "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": 0.004298779065720737
      },
      {
        "cfg": {
          "lr": 0.003
        },
        "mean": 0.0034615940821822733
      },
      {
        "cfg": {
          "lr": 0.01
        },
        "mean": 0.0014865802077110857
      }
    ],
    "full": {
      "mean": 0.0014198302378645167,
      "std": 0.00039839114901047844,
      "per_seed": [
        0.0009327387670055032,
        0.0018641706556081772,
        0.0010785290505737066,
        0.0020708823576569557,
        0.0009260809747502208,
        0.0014055331703275442,
        0.0016443432541564107,
        0.001436363672837615
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.0012014269523206167,
    "std": 0.0003833906235964023,
    "per_seed": [
      0.0005308947875164449,
      0.00110043294262141,
      0.0015973668778315187,
      0.0017763767391443253,
      0.0008540595881640911,
      0.001374695566482842,
      0.000991623499430716,
      0.0013859656173735857
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": -0.00021840328554389998,
    "idea_wins": 7,
    "n_pairs": 8,
    "per_seed_diffs": [
      -0.00040184397948905826,
      -0.0007637377129867673,
      0.000518837827257812,
      -0.00029450561851263046,
      -7.202138658612967e-05,
      -3.0837603844702244e-05,
      -0.0006527197547256947,
      -5.039805546402931e-05
    ],
    "p_value": 0.17895,
    "mde": 0.00033947390618403674,
    "mde_rel_pct": 23.90947150798954,
    "verdict": "no significant win",
    "system_worked": false
  },
  "mechanism_signature": {
    "idea_config": {
      "lr": 0.01
    },
    "idea_sweep": [
      {
        "cfg": {
          "lr": 0.001
        },
        "result": {
          "mean": 0.004815780295757577,
          "std": 0.0014414284088542476,
          "per_seed": [
            0.007190660573542118,
            0.0025714822113513947,
            0.0030915497336536646,
            0.005778642371296883,
            0.005150299984961748,
            0.0042128548957407475,
            0.004502150695770979,
            0.00602860189974308
          ],
          "n": 8
        }
      },
      {
        "cfg": {
          "lr": 0.003
        },
        "result": {
          "mean": 0.0031467917578993365,
          "std": 0.0019972773089779808,
          "per_seed": [
            0.003412304213270545,
            0.002708759158849716,
            0.001344470540061593,
            0.008165908977389336,
            0.0023987311869859695,
            0.002598112216219306,
            0.0016315100947394967,
            0.00291453767567873
          ],
          "n": 8
        }
      },
      {
        "cfg": {
          "lr": 0.01
        },
        "result": {
          "mean": 0.0012014269523206167,
          "std": 0.0003833906235964023,
          "per_seed": [
            0.0005308947875164449,
            0.00110043294262141,
            0.0015973668778315187,
            0.0017763767391443253,
            0.0008540595881640911,
            0.001374695566482842,
            0.000991623499430716,
            0.0013859656173735857
          ],
          "n": 8
        }
      }
    ],
    "prediction": "Hessian response magnitude is finite and tracks curvature scale",
    "rayleigh_q": 5.0501101213740185e-05,
    "observed_hvp_norm": 0.17713873088359833,
    "predicted_curvature_scale": 5.0501101213740185e-05,
    "gradient_norm": 1.5206302404403687,
    "observed_to_predicted_ratio": 3506.9267275331645,
    "confirmed": false
  }
}