AKEW: Assessing Knowledge Editing in the Wild

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Xiaobao, Pan, Liangming, Wang, William Yang, Luu, Anh Tuan
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916430355628032
author Wu, Xiaobao
Pan, Liangming
Wang, William Yang
Luu, Anh Tuan
author_facet Wu, Xiaobao
Pan, Liangming
Wang, William Yang
Luu, Anh Tuan
contents Knowledge editing injects knowledge updates into language models to keep them correct and up-to-date. However, its current evaluations deviate significantly from practice: their knowledge updates solely consist of structured facts derived from meticulously crafted datasets, instead of practical sources -- unstructured texts like news articles, and they often overlook practical real-world knowledge updates. To address these issues, in this paper we propose AKEW (Assessing Knowledge Editing in the Wild), a new practical benchmark for knowledge editing. AKEW fully covers three editing settings of knowledge updates: structured facts, unstructured texts as facts, and extracted triplets. It further introduces new datasets featuring both counterfactual and real-world knowledge updates. Through extensive experiments, we demonstrate the considerable gap between state-of-the-art knowledge-editing methods and practical scenarios. Our analyses further highlight key insights to motivate future research for practical knowledge editing.
format Preprint
id arxiv_https___arxiv_org_abs_2402_18909
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle AKEW: Assessing Knowledge Editing in the Wild
Wu, Xiaobao
Pan, Liangming
Wang, William Yang
Luu, Anh Tuan
Computation and Language
Artificial Intelligence
Knowledge editing injects knowledge updates into language models to keep them correct and up-to-date. However, its current evaluations deviate significantly from practice: their knowledge updates solely consist of structured facts derived from meticulously crafted datasets, instead of practical sources -- unstructured texts like news articles, and they often overlook practical real-world knowledge updates. To address these issues, in this paper we propose AKEW (Assessing Knowledge Editing in the Wild), a new practical benchmark for knowledge editing. AKEW fully covers three editing settings of knowledge updates: structured facts, unstructured texts as facts, and extracted triplets. It further introduces new datasets featuring both counterfactual and real-world knowledge updates. Through extensive experiments, we demonstrate the considerable gap between state-of-the-art knowledge-editing methods and practical scenarios. Our analyses further highlight key insights to motivate future research for practical knowledge editing.
title AKEW: Assessing Knowledge Editing in the Wild
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2402.18909