Hierarchical Orthogonal Residual Spread for Precise Massive Editing in Large Language Models

TL;DR

This paper introduces HORSE, a hierarchical orthogonal residual spread method for large language models that enables precise, stable, and efficient model editing to address safety concerns, outperforming existing approaches.

Contribution

The paper proposes a novel HORSE method focusing on the information matrix's hierarchical orthogonal residual spread, reducing noisy gradients and improving editing stability in LLMs.

Findings

HORSE achieves precise massive editing across diverse scenarios.

Theoretical analysis shows advantages over existing methods.

Extensive experiments validate HORSE's effectiveness on multiple datasets and models.

Abstract

Large language models (LLMs) exhibit exceptional performance across various domains, yet they face critical safety concerns. Model editing has emerged as an effective approach to mitigate these issues. Existing model editing methods often focus on optimizing an information matrix that blends new and old knowledge. While effective, these approaches can be computationally expensive and may cause conflicts. In contrast, we shift our attention to Hierarchical Orthogonal Residual SprEad of the information matrix, which reduces noisy gradients and enables more stable edits from a different perspective. We demonstrate the effectiveness of our method HORSE through a clear theoretical comparison with several popular methods and extensive experiments conducted on two datasets across multiple LLMs. The results show that HORSE maintains precise massive editing across diverse scenarios. The code is available at https://github.com/XiaojieGu/HORSE