{
  "bench_version": 1,
  "track": "planar_ising_autoregressive",
  "model": "mlp_tiny",
  "metric_direction": "lower is better",
  "n_seeds": 8,
  "baseline": {
    "best_cfg": {
      "lr": 0.01,
      "epochs": 5,
      "batch": 128
    },
    "sweep": [
      {
        "cfg": {
          "lr": 0.001,
          "epochs": 5,
          "batch": 128
        },
        "mean": 0.1614346206188202
      },
      {
        "cfg": {
          "lr": 0.003,
          "epochs": 5,
          "batch": 128
        },
        "mean": 0.13007085770368576
      },
      {
        "cfg": {
          "lr": 0.01,
          "epochs": 5,
          "batch": 128
        },
        "mean": 0.12258148938417435
      }
    ],
    "full": {
      "mean": 0.11785224732011557,
      "std": 0.017848149462034908,
      "per_seed": [
        0.15073059499263763,
        0.13608941435813904,
        0.11354892700910568,
        0.08995702117681503,
        0.11521363258361816,
        0.10482487082481384,
        0.12506559491157532,
        0.10738792270421982
      ],
      "n": 8
    }
  },
  "idea": {
    "mean": 0.11573431827127934,
    "std": 0.01748951353530187,
    "per_seed": [
      0.14706209301948547,
      0.1326635181903839,
      0.11457636952400208,
      0.09001504629850388,
      0.10833320766687393,
      0.09629029035568237,
      0.12431652843952179,
      0.11261749267578125
    ],
    "n": 8
  },
  "comparison": {
    "delta_mean": -0.0021179290488362312,
    "idea_wins": 5,
    "n_pairs": 8,
    "per_seed_diffs": [
      -0.0036685019731521606,
      -0.003425896167755127,
      0.0010274425148963928,
      5.8025121688842773e-05,
      -0.006880424916744232,
      -0.00853458046913147,
      -0.0007490664720535278,
      0.005229569971561432
    ],
    "p_value": 0.23005,
    "mde": 0.003718736254444294,
    "mde_rel_pct": 3.1554224369971453,
    "verdict": "no significant win",
    "system_worked": false
  },
  "mechanism_signature": {
    "prediction": "exact soft conditional labels reduce trained-NN conditional probability MSE versus sampled labels",
    "baseline_seed0": {
      "metric": 0.15073059499263763,
      "q_mse": 0.03577027469873428,
      "pred_mean": 0.4622843861579895
    },
    "idea_seed0": {
      "metric": 0.14706209301948547,
      "q_mse": 0.03235262259840965,
      "pred_mean": 0.4577365219593048
    },
    "q_mse_reduction": 0.0034176521003246307,
    "confirmed": true
  },
  "idea_sweep": [
    {
      "cfg": {
        "lr": 0.001,
        "epochs": 5,
        "batch": 128
      },
      "result": {
        "mean": 0.17583313025534153,
        "std": 0.04598349937323497,
        "per_seed": [
          0.15006619691848755,
          0.1609499603509903,
          0.18380002677440643,
          0.15149202942848206,
          0.15694455802440643,
          0.2921749949455261,
          0.17385314404964447,
          0.13738413155078888
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.003,
        "epochs": 5,
        "batch": 128
      },
      "result": {
        "mean": 0.12854224257171154,
        "std": 0.01281968850255869,
        "per_seed": [
          0.1471148431301117,
          0.14259950816631317,
          0.1184714213013649,
          0.11075853556394577,
          0.13586729764938354,
          0.11389961838722229,
          0.136264368891716,
          0.12336234748363495
        ],
        "n": 8
      }
    },
    {
      "cfg": {
        "lr": 0.01,
        "epochs": 5,
        "batch": 128
      },
      "result": {
        "mean": 0.11573431827127934,
        "std": 0.01748951353530187,
        "per_seed": [
          0.14706209301948547,
          0.1326635181903839,
          0.11457636952400208,
          0.09001504629850388,
          0.10833320766687393,
          0.09629029035568237,
          0.12431652843952179,
          0.11261749267578125
        ],
        "n": 8
      }
    }
  ],
  "custom_track": {
    "name": "planar_ising_autoregressive",
    "file": "ising_teacher_track.py",
    "domain": "sequence-level sampling"
  }
}