DeepEdit：知识编辑作为带约束的解码
DeepEdit: Knowledge Editing as Decoding with Constraints
AI总结：
针对大语言模型多步推理中知识编辑的挑战，提出基于深度优先搜索的约束解码框架DEEPEDIT，并构建两个新基准，显著提升推理连贯性与编辑效果。
摘要：
如何编辑多步推理中的知识已成为大语言模型（LLMs）知识编辑（KE）的主要挑战。这一困难源于LLMs在多步推理过程中的幻觉常常导致新知识的错误使用和错误答案。为解决此问题，我们设计了解码约束来“规范”LLMs的推理，在融入新知识时增强逻辑连贯性。我们提出一个新的KE框架：DEEPEDIT（基于深度优先搜索的约束解码用于知识编辑），它通过深度优先搜索增强LLMs生成包含新知识的连贯推理链的能力。我们的搜索选择满足约束的最重要知识作为推理步骤，以有效增加推理深度。除DEEPEDIT外，我们还提出两个新的KE基准：MQUAKE-2002和MQUAKE-HARD，它们对KE方法提供更精确和更具挑战性的评估。定性上，DEEPEDIT使LLMs能够生成涉及新知识的简洁且连贯的推理链。定量上，它在多个KE基准上带来了显著改进。
英文摘要：
How to edit the knowledge in multi-step reasoning has become the major challenge in the knowledge editing (KE) of large language models (LLMs). The difficulty arises because the hallucinations of LLMs during multi-step reasoning often lead to incorrect use of new knowledge and incorrect answers. To address this issue, we design decoding constraints to "regulate" LLMs' reasoning, enhancing logical coherence when incorporating new knowledge. We propose a new KE framework: DEEPEDIT (Depth-first Search-based Constrained Decoding for Knowledge Editing), which enhances LLMs's ability to generate coherent reasoning chains with new knowledge through depth-first search. Our search selects the most important knowledge that satisfies our constraints as the reasoning step to efficiently increase the reasoning depth. In addition to DEEPEDIT, we propose two new KE benchmarks: MQUAKE-2002 and MQUAKE-HARD, which provide more precise and challenging assessments of KE approaches. Qualitatively, DEEPEDIT enables LLMs to produce succinct and coherent reasoning chains involving new knowledge. Quantitatively, it yields significant improvements on multiple KE benchmarks.
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。
