{
  "run_id": "decider__sms_spam__20261011T213723Z",
  "backend": "decider",
  "model_version": "StrandsAgents/strands-decider-2B-hobson-v19@strands-decider-0.1.0",
  "suite_id": "sms_spam",
  "suite_version": "1.0.0",
  "suite_title": "SMS spam",
  "primitive": "noul",
  "protocol": "1.0.0",
  "split": "test",
  "limited": true,
  "limit": 200,
  "prepare_ms": 0.0,
  "latency_p50_ms": 1403.3602565014007,
  "latency_p95_ms": 1996.63814779942,
  "latency_mean_ms": 1531.2698603697936,
  "batch_size": 1,
  "concurrency": 1,
  "cost_usd_total": 0.0,
  "cost_per_1k_usd": 0.0,
  "hardware": {
    "platform": "Linux-6.12.94+-x86_64-with-glibc2.41",
    "python": "3.12.14",
    "machine": "x86_64",
    "processor": "x86_64",
    "cpu_count": 8
  },
  "robustness_probes": {
    "order_reverse": {
      "n": 38,
      "invalid_pairs": 0,
      "flip_rate": 0.0,
      "probability_drift": 0.0,
      "attack_success_rate": null
    },
    "order_shuffle": {
      "n": 38,
      "invalid_pairs": 0,
      "flip_rate": 0.0,
      "probability_drift": 0.0,
      "attack_success_rate": null
    },
    "paraphrase": {
      "n": 38,
      "invalid_pairs": 0,
      "flip_rate": 0.07894736842105263,
      "probability_drift": 0.05488157894736843,
      "attack_success_rate": null
    },
    "irrelevant_context": {
      "n": 38,
      "invalid_pairs": 0,
      "flip_rate": 0.0,
      "probability_drift": 0.03363947368421052,
      "attack_success_rate": null
    },
    "adversarial": {
      "n": 38,
      "invalid_pairs": 0,
      "flip_rate": 0.0,
      "probability_drift": 0.06616315789473684,
      "attack_success_rate": 0.13157894736842105
    }
  },
  "warnings": [],
  "license": "The Hugging Face card lists the license as unknown. This is the public SMS Spam Collection (Almeida, Hidalgo, Yamakami), distributed for spam-filtering research.",
  "n": 200,
  "n_valid": 200,
  "schema_validity": 1.0,
  "accuracy": 1.0,
  "accuracy_ci95": [
    1.0,
    1.0
  ],
  "macro_f1": 1.0,
  "macro_f1_ci95": [
    1.0,
    1.0
  ],
  "ece": 0.1501145,
  "ece_ci95": [
    0.14483142500000004,
    0.15578032500000016
  ],
  "brier": 0.0479441781,
  "brier_ci95": [
    0.043710047617500004,
    0.052731669682499996
  ],
  "brier_max": 2.0,
  "score_mae": null,
  "s1_score": 97.32594564803685,
  "s1_score_ci95": [
    97.21543606708934,
    97.43108614176434
  ],
  "components": {
    "accuracy": 1.0,
    "macro_f1": 1.0,
    "calibration": 0.8498855,
    "brier_quality": 0.97602791095,
    "quality": 0.96518268219,
    "schema_validity": 1.0,
    "robustness": 0.9842105263157894,
    "robustness_measured": true
  },
  "composite_id": "s1-v1",
  "reliability": [
    {
      "bin": 0,
      "low": 0.0,
      "high": 0.1,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 1,
      "low": 0.1,
      "high": 0.2,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 2,
      "low": 0.2,
      "high": 0.3,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 3,
      "low": 0.3,
      "high": 0.4,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 4,
      "low": 0.4,
      "high": 0.5,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 5,
      "low": 0.5,
      "high": 0.6,
      "count": 0,
      "confidence": null,
      "accuracy": null
    },
    {
      "bin": 6,
      "low": 0.6,
      "high": 0.7,
      "count": 3,
      "confidence": 0.6509,
      "accuracy": 1.0
    },
    {
      "bin": 7,
      "low": 0.7,
      "high": 0.8,
      "count": 11,
      "confidence": 0.771809090909091,
      "accuracy": 1.0
    },
    {
      "bin": 8,
      "low": 0.8,
      "high": 0.9,
      "count": 179,
      "confidence": 0.8553346368715083,
      "accuracy": 1.0
    },
    {
      "bin": 9,
      "low": 0.9,
      "high": 1.0,
      "count": 7,
      "confidence": 0.9185142857142857,
      "accuracy": 1.0
    }
  ],
  "selective": [
    {
      "threshold": 0.5,
      "n": 200,
      "coverage": 1.0,
      "risk": 0.0
    },
    {
      "threshold": 0.6,
      "n": 200,
      "coverage": 1.0,
      "risk": 0.0
    },
    {
      "threshold": 0.7,
      "n": 197,
      "coverage": 0.985,
      "risk": 0.0
    },
    {
      "threshold": 0.8,
      "n": 186,
      "coverage": 0.93,
      "risk": 0.0
    },
    {
      "threshold": 0.9,
      "n": 7,
      "coverage": 0.035,
      "risk": 0.0
    },
    {
      "threshold": 0.95,
      "n": 0,
      "coverage": 0.0,
      "risk": null
    }
  ],
  "risk_coverage": [
    {
      "coverage": 0.1,
      "n": 20,
      "risk": 0.0
    },
    {
      "coverage": 0.2,
      "n": 40,
      "risk": 0.0
    },
    {
      "coverage": 0.3,
      "n": 60,
      "risk": 0.0
    },
    {
      "coverage": 0.4,
      "n": 80,
      "risk": 0.0
    },
    {
      "coverage": 0.5,
      "n": 100,
      "risk": 0.0
    },
    {
      "coverage": 0.6,
      "n": 120,
      "risk": 0.0
    },
    {
      "coverage": 0.7,
      "n": 140,
      "risk": 0.0
    },
    {
      "coverage": 0.8,
      "n": 160,
      "risk": 0.0
    },
    {
      "coverage": 0.9,
      "n": 180,
      "risk": 0.0
    },
    {
      "coverage": 1.0,
      "n": 200,
      "risk": 0.0
    }
  ],
  "bootstrap": {
    "n": 1000,
    "seed": 20261006,
    "interval": "percentile 2.5/97.5"
  },
  "notes": {
    "s1_score_interval": "The interval resamples test items and keeps the robustness term fixed at its point estimate. It is not a bootstrap of the probe subset."
  }
}