Editing the Mind of Giants: An In-Depth Exploration of Pitfalls of Knowledge Editing in Large Language Models

TL;DR

This survey critically examines the side effects and challenges of knowledge editing in large language models, emphasizing the need for standardized evaluation metrics and deeper understanding of model knowledge structures.

Contribution

It provides a comprehensive review of knowledge editing pitfalls, introduces consistent evaluation benchmarks, and outlines future research directions in the field.

Findings

Knowledge editing can cause knowledge distortion and degrade model abilities.

Current evaluation metrics are inconsistent and lack standardization.

Deeper understanding of LLMs' internal knowledge structures is necessary.

Abstract

Knowledge editing is a rising technique for efficiently updating factual knowledge in large language models (LLMs) with minimal alteration of parameters. However, recent studies have identified side effects, such as knowledge distortion and the deterioration of general abilities, that have emerged after editing. Despite these findings, evaluating the pitfalls of knowledge editing often relies on inconsistent metrics and benchmarks, lacking a uniform standard. In response, this survey presents a comprehensive study of these side effects, providing a unified perspective on the challenges of knowledge editing in LLMs by conducting experiments with consistent metrics and benchmarks. Additionally, we review related works and outline potential research directions to address these limitations. Our survey highlights the limitations of current knowledge editing methods, emphasizing the need for a deeper understanding of the inner knowledge structures of LLMs and improved knowledge editing methods. To foster future research, we have released the complementary materials publicly in https://github.com/MiuLab/EditLLM-Survey.