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

(универсальный)

Ответ агента:

{
  "worked": true,
  "confidence": 8,
  "verdict": "Natural-parameter information fusion beat the matched arithmetic-fusion recurrent system on the registered dynamics track. Mean test MSE was 0.0159285412 versus 0.0190883005, with 7/8 paired wins and permutation p=0.02315, satisfying the significant-win criterion.",
  "metrics": {
    "baseline": "mean MSE 0.0190883005; std 0.0055838602; best lr 0.006",
    "idea": "mean MSE 0.0159285412; std 0.0046797346; best lr 0.006",
    "paired_delta": "-0.0031597593",
    "p_value": "0.02315",
    "wins": "7/8"
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 bench_experiment.py",
  "files": [
    "bench_experiment.py",
    "bench_report.json"
  ],
  "limitations": "Only the registered dynamics track was tested. The implementation uses learned scalar precisions and fixed complementary masks rather than explicit autodifferentiated Jacobians and noise matrices. Stale-message schedules, graph sweeps, longer rollouts, and quantitative full-Gramian eigenvalue estimation were not tested.",
  "bench_report": {
    "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": 0.5088173821568489
        },
        {
          "cfg": {
            "lr": 0.003
          },
          "mean": 0.06073589529842138
        },
        {
          "cfg": {
            "lr": 0.006
          },
          "mean": 0.023900731466710567
        }
      ],
      "full": {
        "mean": 0.019088300527073443,
        "std": 0.00558386015869873,
        "per_seed": [
          0.018445175141096115,
          0.023440849035978317,
          0.02936750277876854,
          0.024349398910999298,
          0.01488791685551405,
          0.012780136428773403,
          0.01545807532966137,
          0.013977349735796452
        ],
        "n": 8
      }
    },
    "idea": {
      "best_cfg": {
        "lr": 0.006
      },
      "sweep": [
        {
          "cfg": {
            "lr": 0.001
          },
          "mean": 0.3442936111241579
        },
        {
          "cfg": {
            "lr": 0.003
          },
          "mean": 0.03436855808831751
        },
        {
          "cfg": {
            "lr": 0.006
          },
          "mean": 0.015928541193716228
        }
      ],
      "mean": 0.015928541193716228,
      "std": 0.004679734568777004,
      "per_seed": [
        0.0178981926292181,
        0.013649515807628632,
        0.024342793971300125,
        0.021126413717865944,
        0.012977873906493187,
        0.008715411648154259,
        0.01320882048457861,
        0.015509307384490967
      ],
      "n": 8
    },
    "comparison": {
      "delta_mean": -0.003159759333357215,
      "idea_wins": 7,
      "n_pairs": 8,
      "per_seed_diffs": [
        -0.0005469825118780136,
        -0.009791333228349686,
        -0.005024708807468414,
        -0.003222985193133354,
        -0.0019100429490208626,
        -0.0040647247806191444,
        -0.00224925484508276,
        0.0015319576486945152
      ],
      "p_value": 0.02315,
      "verdict": "idea better (significant)",
      "system_worked": true
    },
    "mechanism_signature": {
      "collective_detectability": {
        "window": 2,
        "predicted_lambda_min_positive": true,
        "observed_mean_fused_information": 1.459951639175415,
        "observed_positive_fraction": 1.0,
        "tolerance": "positivity and >0.99 positive fraction",
        "confirmed": true
      }
    }
  },
  "system_verdict": "worked",
  "practical_verdict": "helps",
  "mechanism_ok": 0,
  "system_judged": true
}