Model Editing by Pure Fine-Tuning

arXiv:2402.11078v1 Announce Type: new
Abstract: Fine-tuning is dismissed as not effective for model editing due to its poor performance compared to more specialized methods. However, fine-tuning is simple, agnostic to the architectural details of the model being edited, and able to leverage ongoing advances in standard training methods (e.g., PEFT), making it an appealing choice for a model editor. In this work, we show that pure fine-tuning can be a viable approach to model editing. We propose a slight modification …

abstract advances arxiv cs.ai cs.cl cs.lg editing fine-tuning making peft performance simple standard training type