{
  "model": "deepseek-v4-flash",
  "api_base": "https://uni-api.cstcloud.cn/v1",
  "timestamp": "2026-05-04T17:03:19.718018",
  "total_evaluated": 1200,
  "avg_latency_seconds": 54.19,
  "api_errors": 0,
  "summary": {
    "total_episodes": 1200,
    "detection_accuracy": 0.5858333333333333,
    "miss_rate": 0.3836363636363636,
    "false_alarm_rate": 0.4323076923076923,
    "precision": 0.5467741935483871,
    "recall": 0.6163636363636363,
    "f1_security": 0.5794871794871795,
    "threat_type_accuracy": 0.5811209439528023,
    "parse_failure_rate": 0.0525,
    "breakdown": {
      "TP": {
        "total": 550,
        "correct": 339,
        "accuracy": 0.6163636363636363
      },
      "FP": {
        "total": 400,
        "correct": 216,
        "accuracy": 0.54
      },
      "TN": {
        "total": 250,
        "correct": 148,
        "accuracy": 0.592
      }
    },
    "per_sq": {
      "SQ1": {
        "total": 190,
        "accuracy": 0.6052631578947368,
        "miss_count": 13,
        "false_alarm_count": 60
      },
      "SQ2": {
        "total": 240,
        "accuracy": 0.6375,
        "miss_count": 36,
        "false_alarm_count": 50
      },
      "SQ3": {
        "total": 290,
        "accuracy": 0.5172413793103449,
        "miss_count": 77,
        "false_alarm_count": 62
      },
      "SQ4": {
        "total": 290,
        "accuracy": 0.5655172413793104,
        "miss_count": 49,
        "false_alarm_count": 76
      },
      "SQ5": {
        "total": 190,
        "accuracy": 0.6368421052631579,
        "miss_count": 36,
        "false_alarm_count": 33
      }
    }
  },
  "errors": {
    "total_errors": 639,
    "error_distribution": {
      "FALSE_ALARM": 276,
      "PARSE_FAIL": 35,
      "MISS": 186,
      "WRONG_TYPE": 142
    },
    "error_by_sq": {
      "SQ1": {
        "FALSE_ALARM": 57,
        "PARSE_FAIL": 7,
        "MISS": 11,
        "WRONG_TYPE": 3
      },
      "SQ2": {
        "FALSE_ALARM": 50,
        "PARSE_FAIL": 5,
        "MISS": 32,
        "WRONG_TYPE": 20
      },
      "SQ3": {
        "FALSE_ALARM": 61,
        "PARSE_FAIL": 8,
        "MISS": 71,
        "WRONG_TYPE": 45
      },
      "SQ4": {
        "FALSE_ALARM": 75,
        "PARSE_FAIL": 8,
        "WRONG_TYPE": 49,
        "MISS": 43
      },
      "SQ5": {
        "FALSE_ALARM": 33,
        "MISS": 29,
        "WRONG_TYPE": 25,
        "PARSE_FAIL": 7
      }
    },
    "error_by_category": {
      "device_fault": 14,
      "fire_gas": 40,
      "water_damage": 14,
      "intrusion": 111,
      "behavioral_anomaly": 84,
      "child_specific": 26,
      "elderly_specific": 39
    }
  }
}