arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2403.08319cs.CLcs.AIcs.IRcs.LG
收藏

大语言模型的知识冲突:综述

Knowledge Conflicts for LLMs: A Survey

Westlake University(西湖大学);The Chinese University of Hong Kong(香港中文大学);Tsinghua University(清华大学);University of Cambridge(剑桥大学)

Rongwu Xu, Zehan Qi, Zhijiang Guo, Cunxiang Wang, Hongru Wang, Yue Zhang, Wei Xu

AI总结:

该综述聚焦大语言模型的三类知识冲突,分析其成因、模型表现及现有解决方案,旨在为提升大语言模型鲁棒性、推动相关领域研究提供参考。

摘要:

本综述对大语言模型(LLMs)的知识冲突展开深入分析,重点阐述这类模型在融合上下文知识与参数化知识时面临的复杂挑战。我们聚焦三类知识冲突:上下文-记忆冲突、上下文间冲突与记忆内冲突。这些冲突会显著影响LLMs的可信度与性能,在噪声和错误信息普遍存在的真实应用场景中尤为突出。通过对冲突进行分类、探究其成因、考察LLMs在这类冲突下的行为表现,并梳理现有解决方案,本综述旨在阐明提升LLMs鲁棒性的策略,为推动这一新兴领域的研究提供有价值的参考。

英文摘要:

This survey provides an in-depth analysis of knowledge conflicts for large language models (LLMs), highlighting the complex challenges they encounter when blending contextual and parametric knowledge. Our focus is on three categories of knowledge conflicts: context-memory, inter-context, and intra-memory conflict. These conflicts can significantly impact the trustworthiness and performance of LLMs, especially in real-world applications where noise and misinformation are common. By categorizing these conflicts, exploring the causes, examining the behaviors of LLMs under such conflicts, and reviewing available solutions, this survey aims to shed light on strategies for improving the robustness of LLMs, thereby serving as a valuable resource for advancing research in this evolving area.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑