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

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Mingyang, Lange, Lukas, Adel, Heike, Strötgen, Jannik, Schütze, Hinrich
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916422056148992
author Wang, Mingyang
Lange, Lukas
Adel, Heike
Strötgen, Jannik
Schütze, Hinrich
author_facet Wang, Mingyang
Lange, Lukas
Adel, Heike
Strötgen, Jannik
Schütze, Hinrich
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2410_02433
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization
Wang, Mingyang
Lange, Lukas
Adel, Heike
Strötgen, Jannik
Schütze, Hinrich
Computation and Language
Machine Learning
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.
title Better Call SAUL: Fluent and Consistent Language Model Editing with Generation Regularization
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2410.02433