{
  "model": "Qwen/Qwen3.5-9B",
  "api_base": "http://localhost:8000/v1",
  "timestamp": "2026-05-03T09:31:20.047748",
  "total_evaluated": 1200,
  "avg_latency_seconds": 71.08,
  "api_errors": 1,
  "summary": {
    "total_episodes": 1200,
    "detection_accuracy": 0.5566666666666666,
    "miss_rate": 0.19454545454545455,
    "false_alarm_rate": 0.6430769230769231,
    "precision": 0.5145180023228804,
    "recall": 0.8054545454545454,
    "f1_security": 0.6279234585400425,
    "threat_type_accuracy": 0.40632054176072235,
    "parse_failure_rate": 0.1825,
    "breakdown": {
      "TP": {
        "total": 550,
        "correct": 443,
        "accuracy": 0.8054545454545454
      },
      "FP": {
        "total": 400,
        "correct": 143,
        "accuracy": 0.3575
      },
      "TN": {
        "total": 250,
        "correct": 82,
        "accuracy": 0.328
      }
    },
    "per_sq": {
      "SQ1": {
        "total": 190,
        "accuracy": 0.4842105263157895,
        "miss_count": 18,
        "false_alarm_count": 75
      },
      "SQ2": {
        "total": 240,
        "accuracy": 0.5708333333333333,
        "miss_count": 12,
        "false_alarm_count": 91
      },
      "SQ3": {
        "total": 290,
        "accuracy": 0.5344827586206896,
        "miss_count": 42,
        "false_alarm_count": 93
      },
      "SQ4": {
        "total": 290,
        "accuracy": 0.5586206896551724,
        "miss_count": 16,
        "false_alarm_count": 110
      },
      "SQ5": {
        "total": 190,
        "accuracy": 0.6421052631578947,
        "miss_count": 19,
        "false_alarm_count": 49
      }
    }
  },
  "errors": {
    "total_errors": 795,
    "error_distribution": {
      "FALSE_ALARM": 386,
      "PARSE_FAIL": 91,
      "WRONG_TYPE": 263,
      "MISS": 55
    },
    "error_by_sq": {
      "SQ1": {
        "FALSE_ALARM": 68,
        "PARSE_FAIL": 19,
        "WRONG_TYPE": 3,
        "MISS": 11
      },
      "SQ2": {
        "PARSE_FAIL": 18,
        "FALSE_ALARM": 84,
        "WRONG_TYPE": 56,
        "MISS": 1
      },
      "SQ3": {
        "FALSE_ALARM": 86,
        "PARSE_FAIL": 21,
        "MISS": 28,
        "WRONG_TYPE": 83
      },
      "SQ4": {
        "FALSE_ALARM": 104,
        "PARSE_FAIL": 17,
        "WRONG_TYPE": 69,
        "MISS": 7
      },
      "SQ5": {
        "FALSE_ALARM": 44,
        "PARSE_FAIL": 16,
        "WRONG_TYPE": 52,
        "MISS": 8
      }
    },
    "error_by_category": {
      "device_fault": 14,
      "water_damage": 26,
      "fire_gas": 21,
      "intrusion": 116,
      "behavioral_anomaly": 77,
      "child_specific": 28,
      "elderly_specific": 36
    }
  }
}