FAME: Towards Factual Multi-Task Model Editing

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Zeng, Li, Shan, Yingyu, Liu, Zeming, Yao, Jiashu, Guo, Yuhang
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929549269270528
author Zeng, Li
Shan, Yingyu
Liu, Zeming
Yao, Jiashu
Guo, Yuhang
author_facet Zeng, Li
Shan, Yingyu
Liu, Zeming
Yao, Jiashu
Guo, Yuhang
contents Large language models (LLMs) embed extensive knowledge and utilize it to perform exceptionally well across various tasks. Nevertheless, outdated knowledge or factual errors within LLMs can lead to misleading or incorrect responses, causing significant issues in practical applications. To rectify the fatal flaw without the necessity for costly model retraining, various model editing approaches have been proposed to correct inaccurate knowledge within LLMs in a cost-efficient way. To evaluate these model editing methods, previous work introduced a series of datasets. However, most of the previous datasets only contain fabricated data in a single format, which diverges from real-world model editing scenarios, raising doubts about their usability in practice. To facilitate the application of model editing in real-world scenarios, we propose the challenge of practicality. To resolve such challenges and effectively enhance the capabilities of LLMs, we present FAME, an factual, comprehensive, and multi-task dataset, which is designed to enhance the practicality of model editing. We then propose SKEME, a model editing method that uses a novel caching mechanism to ensure synchronization with the real world. The experiments demonstrate that SKEME performs excellently across various tasks and scenarios, confirming its practicality.
format Preprint
id arxiv_https___arxiv_org_abs_2410_10859
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle FAME: Towards Factual Multi-Task Model Editing
Zeng, Li
Shan, Yingyu
Liu, Zeming
Yao, Jiashu
Guo, Yuhang
Computation and Language
Artificial Intelligence
Large language models (LLMs) embed extensive knowledge and utilize it to perform exceptionally well across various tasks. Nevertheless, outdated knowledge or factual errors within LLMs can lead to misleading or incorrect responses, causing significant issues in practical applications. To rectify the fatal flaw without the necessity for costly model retraining, various model editing approaches have been proposed to correct inaccurate knowledge within LLMs in a cost-efficient way. To evaluate these model editing methods, previous work introduced a series of datasets. However, most of the previous datasets only contain fabricated data in a single format, which diverges from real-world model editing scenarios, raising doubts about their usability in practice. To facilitate the application of model editing in real-world scenarios, we propose the challenge of practicality. To resolve such challenges and effectively enhance the capabilities of LLMs, we present FAME, an factual, comprehensive, and multi-task dataset, which is designed to enhance the practicality of model editing. We then propose SKEME, a model editing method that uses a novel caching mechanism to ensure synchronization with the real world. The experiments demonstrate that SKEME performs excellently across various tasks and scenarios, confirming its practicality.
title FAME: Towards Factual Multi-Task Model Editing
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2410.10859