@inproceedings{lin-xin-ru-etal-2025-yan,
    title = "言行不一:大语言模型决策中的隐性偏见",
    author = "林莘茹, 林莘茹  and
      Li, Luyang  and
      Liu, Xiangting",
    editor = "Sun, Maosong  and
      Duan, Peiyong  and
      Liu, Zhiyuan  and
      Xu, Ruifeng  and
      Sun, Weiwei",
    booktitle = "Proceedings of the 24th {C}hina National Conference on Computational Linguistics ({CCL} 2025)",
    month = aug,
    year = "2025",
    address = "Jinan, China",
    publisher = "Chinese Information Processing Society of China",
    url = "https://preview.aclanthology.org/ingest-ccl/2025.ccl-1.57/",
    pages = "756--768",
    abstract = "``大语言模型的隐性偏见会隐蔽地影响模型的决策过程,使其在应用中难以保证公平性。本文首先构建基于决策的提示数据集进行隐性偏见评估,实验结果表明性能强的大语言模型可能表现出更严重的隐性偏见。进而为了缓解模型的隐性偏见,本文探索了自我反思和模型编辑两类方法。实验发现前者有助于识别隐性偏见,但无法在回答中去偏。在模型编辑实验中通过构建纠偏数据集,得出对模型后四层进行微调可获得最佳去偏效果,这一结论显示出有限参数调整在缓解隐性偏见方面的潜力。''"
}