Forgetting before Learning: Utilizing Parametric Arithmetic for Knowledge Updating in Large Language Models
Preprint
Abstract
Recently Large Language Models (LLMs) have demonstrated their amazing text understanding and generation capabilities. However, even stronger LLMs may still learn incorrect knowledge from the training corpus, as well as some knowledge that is outdated over time. Direct secondary fine-tuning with data containing new knowledge may be ineffective in updating knowledge due to the conflict between old and new knowledge. In this paper, we propose a new paradigm for fine-tuning called F-Learning (Forgetting before Learning), which is based on parametric arithmetic to achieve forgetting of old knowledge and learning of new knowledge. Experimental results on two publicly available datasets demonstrate that our proposed F-Learning can obviously improve the knowledge updating performance of both full fine-tuning and LoRA fine-tuning. Moreover, we have also discovered that forgetting old knowledge by subtracting the parameters of LoRA can achieve a similar effect to subtracting the parameters of full fine-tuning, and sometimes even surpass it significantly.Keywords
All Related Versions
This publication has 5 references indexed in Scilit:
- Knowledge Editing for Large Language Models: A SurveyACM Computing Surveys, 2024
- Editing Large Language Models: Problems, Methods, and Opportunities2023
- Large Language Models with Controllable Working Memory2023
- Plug-and-Play Adaptation for Continuously-updated QA2022
- Zero-Shot Relation Extraction via Reading Comprehension2017