Blame for metadata/label_mapping.json

{
  "label_definitions": {
    "active_fabrication": {
      "consistency": "inconsistent",
      "truth_relation": "contradicts",
      "zh": "主动反向造假"
    },
    "active_full_disclosure": {
      "consistency": "consistent",
      "truth_relation": "matches",
      "zh": "主动全盘坦诚"
    },
    "defensive_counterclaim": {
      "consistency": "inconsistent",
      "truth_relation": "counterclaims",
      "zh": "被动反向辩解"
    },
    "evasive_response": {
      "consistency": "inconsistent",
      "truth_relation": "evades",
      "zh": "消极回避应对"
    },
    "inadvertent_leakage": {
      "consistency": "consistent",
      "truth_relation": "leaks",
      "zh": "无意信息泄露"
    },
    "selective_withholding": {
      "consistency": "consistent",
      "truth_relation": "partial",
      "zh": "主动选择性隐瞒"
    }
  },
  "label_scheme_version": "roomreader_six_class_v1",
  "werewolf_label_map": {
    "active_full_truthful_disclosure": "active_full_disclosure",
    "active_reverse_fabrication": "active_fabrication",
    "defensive_counter_denial": "defensive_counterclaim",
    "evasive_response": "evasive_response",
    "selective_concealment": "selective_withholding",
    "unintentional_information_leakage": "inadvertent_leakage"
  }
}