{
  "model": "deepseek-v4-flash",
  "api_base": "https://uni-api.cstcloud.cn/v1",
  "timestamp": "2026-05-02T01:24:44.383239",
  "total_evaluated": 1200,
  "avg_latency_seconds": 77.93,
  "api_errors": 0,
  "summary": {
    "total_episodes": 1200,
    "detection_accuracy": 0.5766666666666667,
    "miss_rate": 0.5636363636363636,
    "false_alarm_rate": 0.2969230769230769,
    "precision": 0.5542725173210161,
    "recall": 0.43636363636363634,
    "f1_security": 0.48830111902339773,
    "threat_type_accuracy": 0.425,
    "parse_failure_rate": 0.06833333333333333,
    "breakdown": {
      "TP": {
        "total": 550,
        "correct": 240,
        "accuracy": 0.43636363636363634
      },
      "FP": {
        "total": 400,
        "correct": 269,
        "accuracy": 0.6725
      },
      "TN": {
        "total": 250,
        "correct": 183,
        "accuracy": 0.732
      }
    },
    "per_sq": {
      "SQ1": {
        "total": 190,
        "accuracy": 0.5789473684210527,
        "miss_count": 43,
        "false_alarm_count": 37
      },
      "SQ2": {
        "total": 240,
        "accuracy": 0.6291666666666667,
        "miss_count": 52,
        "false_alarm_count": 35
      },
      "SQ3": {
        "total": 290,
        "accuracy": 0.593103448275862,
        "miss_count": 73,
        "false_alarm_count": 44
      },
      "SQ4": {
        "total": 290,
        "accuracy": 0.5310344827586206,
        "miss_count": 81,
        "false_alarm_count": 54
      },
      "SQ5": {
        "total": 190,
        "accuracy": 0.5526315789473685,
        "miss_count": 61,
        "false_alarm_count": 23
      }
    }
  },
  "errors": {
    "total_errors": 646,
    "error_distribution": {
      "FALSE_ALARM": 181,
      "MISS": 284,
      "WRONG_TYPE": 138,
      "PARSE_FAIL": 42,
      "OTHER": 1
    },
    "error_by_sq": {
      "SQ1": {
        "FALSE_ALARM": 37,
        "MISS": 42,
        "WRONG_TYPE": 10,
        "PARSE_FAIL": 1
      },
      "SQ2": {
        "FALSE_ALARM": 29,
        "PARSE_FAIL": 13,
        "MISS": 47,
        "WRONG_TYPE": 33
      },
      "SQ3": {
        "FALSE_ALARM": 41,
        "OTHER": 1,
        "PARSE_FAIL": 5,
        "MISS": 71,
        "WRONG_TYPE": 43
      },
      "SQ4": {
        "FALSE_ALARM": 52,
        "PARSE_FAIL": 15,
        "WRONG_TYPE": 31,
        "MISS": 69
      },
      "SQ5": {
        "FALSE_ALARM": 22,
        "PARSE_FAIL": 8,
        "MISS": 55,
        "WRONG_TYPE": 21
      }
    },
    "error_by_category": {
      "device_fault": 93,
      "fire_gas": 96,
      "water_damage": 44,
      "intrusion": 114,
      "behavioral_anomaly": 36,
      "child_specific": 14,
      "elderly_specific": 25
    }
  }
}