@inproceedings{chengwei-etal-2024-ji,
    title = "基于大模型的交互式谎言识别:数据和模型(Unveiling Lies: Enhancing Large Language Models for Real-World Lie Detection in Interactive Dialogues)",
    author = "Ji, Chengwei  and
      Wang, Siyuan  and
      Li, Taishan  and
      Mou, Xinyi  and
      Zhao, Limin  and
      Xue, Lanqing  and
      Ying, Zhenzhe  and
      Wang, Weiqiang  and
      Huang, Xuanjing  and
      Wei, Zhongyu",
    editor = "Maosong, Sun  and
      Jiye, Liang  and
      Xianpei, Han  and
      Zhiyuan, Liu  and
      Yulan, He",
    booktitle = "Proceedings of the 23rd Chinese National Conference on Computational Linguistics (Volume 1: Main Conference)",
    month = jul,
    year = "2024",
    address = "Taiyuan, China",
    publisher = "Chinese Information Processing Society of China",
    url = "https://preview.aclanthology.org/name-variant-aaron-steven-white/2024.ccl-1.68/",
    pages = "870--882",
    language = "zho",
    abstract = "``面向对话交互过程的谎言识别技术在不同的应用场景有广泛的应用需求。现有的鉴谎技术往往在整体的对话级别上给出最终决策,而缺乏对细粒度谎言特征和线索的逻辑分析,难以满足场景中对于可解释性的需求。本文提出了谎言指征和语义不一致线索的概念,用于帮助识别对话中的谎言,提升鉴谎方法的可解释性。文章同时提出一个谎言识别框架,用于训练谎言识别大语言模型(LD-LLM)。它利用细粒度的谎言指征并且发现对话中是否存在语义不一致线索,以实现更可靠的谎言识别。文章在真实交互场景中构建了两个谎言识别数据集FinLIE和IDLIE,分别关注金融风控场景和身份识别场景。实验结果表明,基于这两个数据集创建的指令数据集微调得到的LD-LLM,在基于真实交互的谎言识别上达到了最先进的水平。''"
}