# Стенд-проверка (stage-2) · промт оператора:

(универсальный)

Ответ агента:

{
  "worked": false,
  "confidence": 9,
  "verdict": "Implemented Energy-Riesz checkpoint selection on the structurally matched built-in poisson_boundary PDE track using shared mlp_tiny and Adam training. The trained-model signature confirmed monotone nested auxiliary scores, but held-out MSE was significantly worse: delta_mean=+0.019328 and permutation p=0.01485. The mechanism transferred mathematically, but not as a practical benchmark improvement.",
  "metrics": {
    "baseline": "poisson_boundary/mlp_tiny, Adam lr=0.01, 24 epochs: mean test MSE 0.0438708, std 0.0028396",
    "idea": "Best selector setting lr=0.01, auxiliary level=2: mean test MSE 0.0631988, std 0.0111461; paired delta +0.019328, p=0.01485; 0/8 wins",
    "mechanism_signature": {
      "nested_scores": {
        "level_2": 3.402205,
        "level_3": 3.536085,
        "level_4": 3.538054
      },
      "confirmed": true
    }
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 bench_energy_riesz.py",
  "files": [
    "bench_energy_riesz.py",
    "poisson_track.py",
    "bench_report.json"
  ],
  "limitations": "Tested one built-in PDE track, one mlp_tiny architecture, 24 epochs, 400 training and 400 test samples, three learning rates, and three auxiliary levels. No larger models, longer schedules, PINN collocation loss, or quadrature-resolution study was run.",
  "bench_report": {
    "bench_version": 1,
    "track": "poisson_boundary",
    "model": "mlp_tiny",
    "metric_direction": "lower is better",
    "n_seeds": 8,
    "baseline": {
      "best_cfg": {
        "lr": 0.01
      },
      "sweep": [
        {
          "cfg": {
            "lr": 0.001
          },
          "mean": 0.0687526818
        },
        {
          "cfg": {
            "lr": 0.003
          },
          "mean": 0.0577650378
        },
        {
          "cfg": {
            "lr": 0.01
          },
          "mean": 0.0432244567
        }
      ],
      "full": {
        "mean": 0.0438708267,
        "std": 0.0028395563,
        "per_seed": [
          0.0402809083,
          0.0446168296,
          0.0446245745,
          0.0433755144,
          0.038494233,
          0.0466628,
          0.0467552878,
          0.046156466
        ],
        "n": 8
      }
    },
    "idea": {
      "mean": 0.0631987923,
      "std": 0.0111460654,
      "per_seed": [
        0.0402809083,
        0.0720510557,
        0.0720613971,
        0.0550173223,
        0.0545946844,
        0.0707400963,
        0.0726064667,
        0.0682384074
      ],
      "n": 8
    },
    "comparison": {
      "delta_mean": 0.0193279656,
      "idea_wins": 0,
      "n_pairs": 8,
      "p_value": 0.01485,
      "verdict": "idea worse (significant)",
      "system_worked": false
    },
    "mechanism_signature": {
      "nested_probe": {
        "2": 3.4022051939,
        "3": 3.5360848178,
        "4": 3.538053974
      },
      "predicted_monotone": true,
      "confirmed": true
    },
    "runtime_seconds": 117.34
  },
  "system_verdict": "partial",
  "practical_verdict": "harms",
  "mechanism_ok": 1,
  "system_judged": true
}