研究大模型在政治问答中如何处理已知与未知信息，发现其纠错能力有限。
Can LLMs Ground when they (Don't) Know: A Study on Direct and Loaded Political Questions
- 对比直接提问与含误导前提的问题，测试模型接地能力
- 多数情况下模型未能纠正用户错误信念，尤其在政治敏感议题上
- 适用于关注AI misinformation风险的研究者与政策制定者
人类交流依赖对话接地，使对话双方即使知识不完全也能达成理解。本文研究大语言模型（LLMs）在政治领域面对已知与未知信息时的接地能力，聚焦于易产生虚假信息和接地失败的情境。我们评估模型回答直接知识性问题和预设错误前提的诱导性问题的表现，考察其是否能主动进行接地并纠正用户错误信念，以及这种能力如何受模型知识水平和政治偏见影响。研究发现，模型在处理错误信念时存在显著缺陷，难以有效纠正用户误解，尤其在政治议题上表现更差，凸显其在缓解政治话语中虚假信息传播方面的局限性。
原文摘要 · Abstract (English)
Communication among humans relies on conversational grounding, allowing interlocutors to reach mutual understanding even when they do not have perfect knowledge and must resolve discrepancies in each other's beliefs. This paper investigates how large language models (LLMs) manage common ground in cases where they (don't) possess knowledge, focusing on facts in the political domain where the risk of misinformation and grounding failure is high. We examine the ability of LLMs to answer direct knowledge questions and loaded questions that presuppose misinformation. We evaluate whether loaded questions lead LLMs to engage in active grounding and correct false user beliefs, in connection to their level of knowledge and their political bias. Our findings highlight significant challenges in LLMs' ability to engage in grounding and reject false user beliefs, raising concerns about their role in mitigating misinformation in political discourse.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。