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

(универсальный)

Ответ агента:

{
  "worked": false,
  "confidence": 10,
  "verdict": "Ran the registered geometry_support_regression benchmark with matched MLP regressors, a tuned baseline sweep, and 8 paired seeds. The baseline achieved mean MSE 0.0503475652 at lr=0.006 versus 0.0729937740 for the idea; paired delta was +0.0226462088 with permutation p=0.0081, so the idea was significantly worse. The trained-model mechanism signature was not confirmed because observed mean radius-squared was 2.61810413 rather than the predicted 5.0.",
  "metrics": {
    "baseline": "Best lr=0.006; sweep means were 0.0730441455 at lr=0.001, 0.0594160510 at lr=0.003, and 0.0495788129 at lr=0.006; full 8-seed mean MSE=0.0503475652.",
    "idea": "Full 8-seed mean MSE=0.0729937740; paired delta=+0.0226462088; permutation p=0.0081; idea wins=0/8."
  },
  "bench_report": {
    "bench_version": 1,
    "track": "geometry_support_regression",
    "model": "matched_mlp_embedding_regressor",
    "metric_direction": "lower is better",
    "n_seeds": 8,
    "baseline": {
      "best_cfg": {
        "lr": 0.006
      },
      "sweep": [
        {
          "cfg": {
            "lr": 0.001
          },
          "mean": 0.07304414547979832
        },
        {
          "cfg": {
            "lr": 0.003
          },
          "mean": 0.059416051022708416
        },
        {
          "cfg": {
            "lr": 0.006
          },
          "mean": 0.04957881290465593
        }
      ],
      "full": {
        "mean": 0.050347565207630396,
        "std": 0.007310327369941531,
        "per_seed": [
          0.052693966776132584,
          0.044472310692071915,
          0.04606407880783081,
          0.055084895342588425,
          0.06478562951087952,
          0.03895844891667366,
          0.04829391837120056,
          0.052427273243665695
        ],
        "n": 8
      }
    },
    "idea": {
      "mean": 0.07299377396702766,
      "std": 0.005527255691171378,
      "per_seed": [
        0.07619384676218033,
        0.0728721097111702,
        0.07278692722320557,
        0.07364533096551895,
        0.07819892466068268,
        0.06470585614442825,
        0.06447294354438782,
        0.08107425272464752
      ],
      "n": 8
    },
    "comparison": {
      "delta_mean": 0.02264620875939727,
      "idea_wins": 0,
      "n_pairs": 8,
      "per_seed_diffs": [
        0.023499879986047745,
        0.028399799019098282,
        0.026722848415374756,
        0.013413295149803162,
        0.025747407227754593,
        0.016179025173187256,
        0.028646979480981827
      ],
      "p_value": 0.0081,
      "mde": 0.004894412730734753,
      "mde_rel_pct": 9.721250095313414,
      "verdict": "idea worse (significant)",
      "system_worked": false
    },
    "mechanism_signature": {
      "prediction": "calibrated compact density predicts nominal E[r2]=d=5",
      "predicted_mean_radius2": 5.0,
      "observed_mean_radius2_on_trained_models": 2.6181041300296783,
      "absolute_error": 2.3818958699703217,
      "mean_inside_support_fraction": 0.9849999696016312,
      "confirmed": false
    }
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 stage2_compact_bench.py",
  "files": [
    "stage2_compact_bench.py",
    "bench_report.json"
  ],
  "limitations": "Used the registered geometry-support regression track rather than real image OOD detection. CIFAR-10/MVTec, robust finite-sample covariance estimation, pretrained encoders, AUROC, and deployment-time anomaly detection were not tested.",
  "system_verdict": "failed",
  "practical_verdict": "harms",
  "mechanism_ok": 0,
  "system_judged": true
}