{
  "model": "Qwen/Qwen3.5-9B",
  "api_base": "http://localhost:8000/v1",
  "timestamp": "2026-05-04T17:41:47.465933",
  "total_evaluated": 60,
  "summary": {
    "total_episodes": 60,
    "detection_accuracy": 0.5833333333333334,
    "miss_rate": 0.39285714285714285,
    "false_alarm_rate": 0.4375,
    "precision": 0.5483870967741935,
    "recall": 0.6071428571428571,
    "f1_security": 0.5762711864406779,
    "threat_type_accuracy": 0.35294117647058826,
    "parse_failure_rate": 0.0,
    "breakdown": {
      "TP": {
        "total": 28,
        "correct": 17,
        "accuracy": 0.6071428571428571
      },
      "FP": {
        "total": 28,
        "correct": 15,
        "accuracy": 0.5357142857142857
      },
      "TN": {
        "total": 4,
        "correct": 3,
        "accuracy": 0.75
      }
    },
    "per_sq": {
      "SQ1": {
        "total": 13,
        "accuracy": 0.6153846153846154,
        "miss_count": 3,
        "false_alarm_count": 2
      },
      "SQ2": {
        "total": 6,
        "accuracy": 0.8333333333333334,
        "miss_count": 0,
        "false_alarm_count": 1
      },
      "SQ3": {
        "total": 17,
        "accuracy": 0.6470588235294118,
        "miss_count": 3,
        "false_alarm_count": 3
      },
      "SQ4": {
        "total": 17,
        "accuracy": 0.4117647058823529,
        "miss_count": 5,
        "false_alarm_count": 5
      },
      "SQ5": {
        "total": 7,
        "accuracy": 0.5714285714285714,
        "miss_count": 0,
        "false_alarm_count": 3
      }
    }
  },
  "errors": {
    "total_errors": 36,
    "error_distribution": {
      "FALSE_ALARM": 14,
      "MISS": 11,
      "WRONG_TYPE": 11
    },
    "error_by_sq": {
      "SQ1": {
        "FALSE_ALARM": 2,
        "MISS": 3
      },
      "SQ2": {
        "FALSE_ALARM": 1,
        "WRONG_TYPE": 3
      },
      "SQ3": {
        "FALSE_ALARM": 3,
        "WRONG_TYPE": 5,
        "MISS": 3
      },
      "SQ4": {
        "FALSE_ALARM": 5,
        "MISS": 5,
        "WRONG_TYPE": 2
      },
      "SQ5": {
        "FALSE_ALARM": 3,
        "WRONG_TYPE": 1
      }
    },
    "error_by_category": {
      "device_fault": 3,
      "intrusion": 3,
      "water_damage": 2,
      "fire_gas": 2,
      "behavioral_anomaly": 4,
      "child_specific": 4,
      "elderly_specific": 4
    }
  },
  "pipeline": "EGPv2"
}