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

(универсальный)

Ответ агента:

{
  "worked": false,
  "confidence": 10,
  "verdict": "Implemented and evaluated the adaptive CBF layer around the shared rnn_small model using bench.train_model, paired seeds, baseline tuning, and a trained-model mechanism signature. The residual-tracking prediction held quantitatively, but adaptive filtering significantly worsened test MSE versus tuned static clipping, so the idea did not win.",
  "metrics": {
    "baseline": "Mean test MSE 0.0046574123844038695; best configuration lr=0.01, clip_margin=0.0.",
    "idea": "Mean test MSE 0.0047074658796191216; best configuration lr=0.01, kappa=0.5; paired delta +0.00005005349521325204; p=0.0081; 0/8 wins.",
    "mechanism_signature": "Predicted-versus-observed residual MAE 0.000017471262253820896, relative MAE 0.007133921903418612, confirmed=true."
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 stage2_adaptive_cbf_bench.py",
  "files": [
    "stage2_adaptive_cbf_bench.py",
    "bench_report.json"
  ],
  "limitations": "The registered dynamics task is supervised pendulum rollout prediction rather than interactive RL, so closed-loop safety violations, return, and intervention-versus-safety Pareto fronts were not measured. Multidimensional QPs, online model retraining, stochastic disturbances, and actuator/environment closed-loop evaluation 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": 0.01,
        "clip_margin": 0.0
      },
      "sweep": [
        {
          "cfg": {
            "lr": 0.001,
            "clip_margin": 0.0
          },
          "mean": 0.008678926387801766
        },
        {
          "cfg": {
            "lr": 0.001,
            "clip_margin": 0.03
          },
          "mean": 0.010034722508862615
        },
        {
          "cfg": {
            "lr": 0.001,
            "clip_margin": 0.08
          },
          "mean": 0.012918951804749668
        },
        {
          "cfg": {
            "lr": 0.003,
            "clip_margin": 0.0
          },
          "mean": 0.006641886895522475
        },
        {
          "cfg": {
            "lr": 0.003,
            "clip_margin": 0.03
          },
          "mean": 0.006899000000000001
        },
        {
          "cfg": {
            "lr": 0.003,
            "clip_margin": 0.08
          },
          "mean": 0.008067000000000001
        },
        {
          "cfg": {
            "lr": 0.01,
            "clip_margin": 0.0
          },
          "mean": 0.005315697111655027
        },
        {
          "cfg": {
            "lr": 0.01,
            "clip_margin": 0.03
          },
          "mean": 0.006531292689032853
        },
        {
          "cfg": {
            "lr": 0.01,
            "clip_margin": 0.08
          },
          "mean": 0.008570412872359157
        }
      ],
      "full": {
        "mean": 0.0046574123844038695,
        "std": 0.0013966285668353233,
        "per_seed": [
          0.003739116946235299,
          0.004545919597148895,
          0.005411764141172171,
          0.007565987762063742,
          0.004889188800007105,
          0.002318228594958782,
          0.00456235371530056,
          0.004226739518344402
        ],
        "n": 8
      }
    },
    "idea": {
      "mean": 0.0047074658796191216,
      "std": 0.0014044730096107532,
      "per_seed": [
        0.003767308546230197,
        0.004722419194877148,
        0.005457298830151558,
        0.007620518561452627,
        0.004916135687381029,
        0.002345199463889003,
        0.004579263739287853,
        0.0042515830136835575
      ],
      "n": 8
    },
    "comparison": {
      "delta_mean": 5.005349521325204e-05,
      "idea_wins": 0,
      "n_pairs": 8,
      "per_seed_diffs": [
        2.819159999489784e-05,
        0.0001764995977282524,
        4.553468897938728e-05,
        5.45307993888855e-05,
        2.6946887373924255e-05,
        2.6970868930220604e-05,
        1.6910023983493806e-05,
        2.4843495339155197e-05
      ],
      "p_value": 0.0081,
      "mde": 4.391403299454989e-05,
      "mde_rel_pct": 0.9428847903098175,
      "verdict": "idea worse (significant)",
      "system_worked": false
    },
    "mechanism_signature": {
      "selected_idea_cfg": {
        "lr": 0.01,
        "kappa": 0.5
      },
      "baseline_best_cfg": {
        "lr": 0.01,
        "clip_margin": 0.0
      },
      "prediction": "adaptive ebar tracks observed model residual and larger kappa increases intervention",
      "predicted_vs_observed_mae": 1.7471262253820896e-05,
      "relative_mae": 0.007133921903418612,
      "confirmed": true
    }
  },
  "system_verdict": "partial",
  "practical_verdict": "harms",
  "mechanism_ok": 1,
  "system_judged": true
}