Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization

TL;DR

SAUL is a novel model editing approach that uses sentence concatenation with augmented facts to improve editing efficiency and reliability while preserving generation quality and reducing computational costs.

Contribution

The paper introduces SAUL, a new model editing method that employs generation regularization through sentence concatenation with augmented facts, addressing limitations of existing methods.

Findings

SAUL outperforms state-of-the-art methods on three benchmarks.

SAUL maintains generation fluency and consistency after edits.

SAUL reduces computational overhead compared to previous approaches.

Abstract

To ensure large language models contain up-to-date knowledge, they need to be updated regularly. However, model editing is challenging as it might also affect knowledge that is unrelated to the new data. State-of-the-art methods identify parameters associated with specific knowledge and then modify them via direct weight updates. However, these locate-and-edit methods suffer from heavy computational overhead and lack theoretical validation. In contrast, directly fine-tuning the model on requested edits affects the model's behavior on unrelated knowledge, and significantly damages the model's generation fluency and consistency. To address these challenges, we propose SAUL, a streamlined model editing method that uses sentence concatenation with augmented random facts for generation regularization. Evaluations on three model editing benchmarks show that SAUL is a practical and reliable solution for model editing outperforming state-of-the-art methods while maintaining generation quality and reducing computational overhead.