Learning to Edit Knowledge via Instruction-based Chain-of-Thought Prompting

TL;DR

This paper introduces CoT2Edit, a novel method enabling large language models to effectively edit and utilize both structured and unstructured knowledge through reasoning and retrieval, improving generalization in knowledge editing tasks.

Contribution

It proposes a new paradigm using Chain of Thought reasoning for knowledge editing, incorporating supervised fine-tuning and retrieval-augmented generation to enhance model performance.

Findings

Achieves strong generalization across six diverse knowledge editing scenarios.

Uses a single round of training on three open-source models.

Code is publicly available at the provided GitHub link.

Abstract

Large language models (LLMs) can effectively handle outdated information through knowledge editing. However, current approaches face two key limitations: (I) Poor generalization: Most approaches rigidly inject new knowledge without ensuring that the model can use it effectively to solve practical problems. (II) Narrow scope: Current methods focus primarily on structured fact triples, overlooking the diverse unstructured forms of factual information (e.g., news, articles) prevalent in real-world contexts. To address these challenges, we propose a new paradigm: teaching LLMs to edit knowledge via Chain of Thoughts (CoTs) reasoning (CoT2Edit). We first leverage language model agents for both structured and unstructured edited data to generate CoTs, building high-quality instruction data. The model is then trained to reason over edited knowledge through supervised fine-tuning (SFT) and Group Relative Policy Optimization (GRPO). At inference time, we integrate Retrieval-Augmented Generation (RAG) to dynamically retrieve relevant edited facts for real-time knowledge editing. Experimental results demonstrate that our method achieves strong generalization across six diverse knowledge editing scenarios with just a single round of training on three open-source language models. The codes are available at https://github.com/FredJDean/CoT2Edit.