{
  "base": {
    "n": 36,
    "valid_json_rate": 0.6388888888888888,
    "strict_json_rate": 0.0,
    "category_accuracy": 0.2777777777777778,
    "category_macro_f1": 0.2899551066217733,
    "priority_accuracy": 0.08333333333333333,
    "priority_macro_f1": 0.10323886639676114,
    "sentiment_accuracy": 0.3055555555555556,
    "sentiment_macro_f1": 0.3955067920585162,
    "recommended_action_accuracy": 0.1111111111111111,
    "recommended_action_macro_f1": 0.11929824561403508,
    "escalation_recall": 0.0,
    "escalation_precision": null,
    "exact_labels_accuracy": 0.0,
    "latency_p50_seconds": 4.410080273000062,
    "latency_p95_seconds": 6.969468082999924
  },
  "ft": {
    "n": 36,
    "valid_json_rate": 1.0,
    "strict_json_rate": 1.0,
    "category_accuracy": 0.8055555555555556,
    "category_macro_f1": 0.8024691358024693,
    "priority_accuracy": 0.6944444444444444,
    "priority_macro_f1": 0.7069444444444444,
    "sentiment_accuracy": 0.9444444444444444,
    "sentiment_macro_f1": 0.9036890645586299,
    "recommended_action_accuracy": 0.6666666666666666,
    "recommended_action_macro_f1": 0.7038888888888888,
    "escalation_recall": 0.47058823529411764,
    "escalation_precision": 1.0,
    "exact_labels_accuracy": 0.4444444444444444,
    "latency_p50_seconds": 7.1500980320000735,
    "latency_p95_seconds": 10.085614330999988
  },
  "delta": {
    "valid_json_rate": 0.36111111111111116,
    "strict_json_rate": 1.0,
    "category_accuracy": 0.5277777777777778,
    "category_macro_f1": 0.512514029180696,
    "priority_accuracy": 0.611111111111111,
    "priority_macro_f1": 0.6037055780476832,
    "sentiment_accuracy": 0.6388888888888888,
    "sentiment_macro_f1": 0.5081822725001137,
    "recommended_action_accuracy": 0.5555555555555556,
    "recommended_action_macro_f1": 0.5845906432748537,
    "escalation_recall": 0.47058823529411764,
    "exact_labels_accuracy": 0.4444444444444444,
    "latency_p50_seconds": 2.740017759000011,
    "latency_p95_seconds": 3.1161462480000637
  },
  "category_macro_f1_delta_ci95": [
    0.3272246272246273,
    0.6654768233715601
  ],
  "bootstrap_samples": 1000,
  "seed": 42,
  "hard_cases": {
    "base": {
      "n": 19,
      "valid_json_rate": 0.5263157894736842,
      "strict_json_rate": 0.0,
      "category_accuracy": 0.2631578947368421,
      "category_macro_f1": 0.27248677248677244,
      "priority_accuracy": 0.05263157894736842,
      "priority_macro_f1": 0.041666666666666664,
      "sentiment_accuracy": 0.15789473684210525,
      "sentiment_macro_f1": 0.125,
      "recommended_action_accuracy": 0.0,
      "recommended_action_macro_f1": 0.0,
      "escalation_recall": 0.0,
      "escalation_precision": null,
      "exact_labels_accuracy": 0.0,
      "latency_p50_seconds": 4.480382373999987,
      "latency_p95_seconds": 8.558520648000012
    },
    "ft": {
      "n": 19,
      "valid_json_rate": 1.0,
      "strict_json_rate": 1.0,
      "category_accuracy": 0.6842105263157895,
      "category_macro_f1": 0.5944444444444446,
      "priority_accuracy": 0.6842105263157895,
      "priority_macro_f1": 0.5446428571428571,
      "sentiment_accuracy": 0.9473684210526315,
      "sentiment_macro_f1": 0.6533333333333333,
      "recommended_action_accuracy": 0.5263157894736842,
      "recommended_action_macro_f1": 0.3844444444444444,
      "escalation_recall": 0.47058823529411764,
      "escalation_precision": 1.0,
      "exact_labels_accuracy": 0.3684210526315789,
      "latency_p50_seconds": 7.478730603000031,
      "latency_p95_seconds": 10.085511394999912
    }
  }
}