EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models

Large Language Models (LLMs) usually suffer from knowledge cutoff or fallacy issues, which means they are unaware of unseen events or generate text with incorrect facts owing to outdated/noisy data. To this end, many knowledge editing approaches for LLMs have emerged -- aiming to subtly inject/edit...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Wang, Peng, Zhang, Ningyu, Tian, Bozhong, Xi, Zekun, Yao, Yunzhi, Xu, Ziwen, Wang, Mengru, Mao, Shengyu, Wang, Xiaohan, Cheng, Siyuan, Liu, Kangwei, Ni, Yuansheng, Zheng, Guozhou, Chen, Huajun
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Wang, Peng
Zhang, Ningyu
Tian, Bozhong
Xi, Zekun
Yao, Yunzhi
Xu, Ziwen
Wang, Mengru
Mao, Shengyu
Wang, Xiaohan
Cheng, Siyuan
Liu, Kangwei
Ni, Yuansheng
Zheng, Guozhou
Chen, Huajun
description Large Language Models (LLMs) usually suffer from knowledge cutoff or fallacy issues, which means they are unaware of unseen events or generate text with incorrect facts owing to outdated/noisy data. To this end, many knowledge editing approaches for LLMs have emerged -- aiming to subtly inject/edit updated knowledge or adjust undesired behavior while minimizing the impact on unrelated inputs. Nevertheless, due to significant differences among various knowledge editing methods and the variations in task setups, there is no standard implementation framework available for the community, which hinders practitioners from applying knowledge editing to applications. To address these issues, we propose EasyEdit, an easy-to-use knowledge editing framework for LLMs. It supports various cutting-edge knowledge editing approaches and can be readily applied to many well-known LLMs such as T5, GPT-J, LlaMA, etc. Empirically, we report the knowledge editing results on LlaMA-2 with EasyEdit, demonstrating that knowledge editing surpasses traditional fine-tuning in terms of reliability and generalization. We have released the source code on GitHub, along with Google Colab tutorials and comprehensive documentation for beginners to get started. Besides, we present an online system for real-time knowledge editing, and a demo video.
doi_str_mv 10.48550/arxiv.2308.07269
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2308_07269</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2308_07269</sourcerecordid><originalsourceid>FETCH-LOGICAL-a679-a9f1b50b8cd9713662c37a2b76ef990bd51698beb7ed44cf5861432e327c1f803</originalsourceid><addsrcrecordid>eNotj8tOwzAURL1hgQofwAr_gIMf8YtdVaVQEWDTfXQd21FEGiOnpfTv2xRWM6MjjXQQemC0KI2U9Anyb_9TcEFNQTVX9hZ9VDCdKt_vn_FyxPMg-0QOU8BvYzoOwXcBz7gfO7zOsAvHlL9wTBnXkC-shrE7wKW8Jx-G6Q7dRBimcP-fC7RdV9vVK6k_XzarZU1AaUvARuYkdab1VjOhFG-FBu60CtFa6rxkyhoXnA6-LNsojWKl4EFw3bJoqFigx7_bq1Dznfsd5FMzizVXMXEG6mJHzw</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models</title><source>arXiv.org</source><creator>Wang, Peng ; Zhang, Ningyu ; Tian, Bozhong ; Xi, Zekun ; Yao, Yunzhi ; Xu, Ziwen ; Wang, Mengru ; Mao, Shengyu ; Wang, Xiaohan ; Cheng, Siyuan ; Liu, Kangwei ; Ni, Yuansheng ; Zheng, Guozhou ; Chen, Huajun</creator><creatorcontrib>Wang, Peng ; Zhang, Ningyu ; Tian, Bozhong ; Xi, Zekun ; Yao, Yunzhi ; Xu, Ziwen ; Wang, Mengru ; Mao, Shengyu ; Wang, Xiaohan ; Cheng, Siyuan ; Liu, Kangwei ; Ni, Yuansheng ; Zheng, Guozhou ; Chen, Huajun</creatorcontrib><description>Large Language Models (LLMs) usually suffer from knowledge cutoff or fallacy issues, which means they are unaware of unseen events or generate text with incorrect facts owing to outdated/noisy data. To this end, many knowledge editing approaches for LLMs have emerged -- aiming to subtly inject/edit updated knowledge or adjust undesired behavior while minimizing the impact on unrelated inputs. Nevertheless, due to significant differences among various knowledge editing methods and the variations in task setups, there is no standard implementation framework available for the community, which hinders practitioners from applying knowledge editing to applications. To address these issues, we propose EasyEdit, an easy-to-use knowledge editing framework for LLMs. It supports various cutting-edge knowledge editing approaches and can be readily applied to many well-known LLMs such as T5, GPT-J, LlaMA, etc. Empirically, we report the knowledge editing results on LlaMA-2 with EasyEdit, demonstrating that knowledge editing surpasses traditional fine-tuning in terms of reliability and generalization. We have released the source code on GitHub, along with Google Colab tutorials and comprehensive documentation for beginners to get started. Besides, we present an online system for real-time knowledge editing, and a demo video.</description><identifier>DOI: 10.48550/arxiv.2308.07269</identifier><language>eng</language><subject>Computer Science - Artificial Intelligence ; Computer Science - Computation and Language ; Computer Science - Computer Vision and Pattern Recognition ; Computer Science - Information Retrieval ; Computer Science - Learning</subject><creationdate>2023-08</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,781,886</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2308.07269$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2308.07269$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Wang, Peng</creatorcontrib><creatorcontrib>Zhang, Ningyu</creatorcontrib><creatorcontrib>Tian, Bozhong</creatorcontrib><creatorcontrib>Xi, Zekun</creatorcontrib><creatorcontrib>Yao, Yunzhi</creatorcontrib><creatorcontrib>Xu, Ziwen</creatorcontrib><creatorcontrib>Wang, Mengru</creatorcontrib><creatorcontrib>Mao, Shengyu</creatorcontrib><creatorcontrib>Wang, Xiaohan</creatorcontrib><creatorcontrib>Cheng, Siyuan</creatorcontrib><creatorcontrib>Liu, Kangwei</creatorcontrib><creatorcontrib>Ni, Yuansheng</creatorcontrib><creatorcontrib>Zheng, Guozhou</creatorcontrib><creatorcontrib>Chen, Huajun</creatorcontrib><title>EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models</title><description>Large Language Models (LLMs) usually suffer from knowledge cutoff or fallacy issues, which means they are unaware of unseen events or generate text with incorrect facts owing to outdated/noisy data. To this end, many knowledge editing approaches for LLMs have emerged -- aiming to subtly inject/edit updated knowledge or adjust undesired behavior while minimizing the impact on unrelated inputs. Nevertheless, due to significant differences among various knowledge editing methods and the variations in task setups, there is no standard implementation framework available for the community, which hinders practitioners from applying knowledge editing to applications. To address these issues, we propose EasyEdit, an easy-to-use knowledge editing framework for LLMs. It supports various cutting-edge knowledge editing approaches and can be readily applied to many well-known LLMs such as T5, GPT-J, LlaMA, etc. Empirically, we report the knowledge editing results on LlaMA-2 with EasyEdit, demonstrating that knowledge editing surpasses traditional fine-tuning in terms of reliability and generalization. We have released the source code on GitHub, along with Google Colab tutorials and comprehensive documentation for beginners to get started. Besides, we present an online system for real-time knowledge editing, and a demo video.</description><subject>Computer Science - Artificial Intelligence</subject><subject>Computer Science - Computation and Language</subject><subject>Computer Science - Computer Vision and Pattern Recognition</subject><subject>Computer Science - Information Retrieval</subject><subject>Computer Science - Learning</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2023</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj8tOwzAURL1hgQofwAr_gIMf8YtdVaVQEWDTfXQd21FEGiOnpfTv2xRWM6MjjXQQemC0KI2U9Anyb_9TcEFNQTVX9hZ9VDCdKt_vn_FyxPMg-0QOU8BvYzoOwXcBz7gfO7zOsAvHlL9wTBnXkC-shrE7wKW8Jx-G6Q7dRBimcP-fC7RdV9vVK6k_XzarZU1AaUvARuYkdab1VjOhFG-FBu60CtFa6rxkyhoXnA6-LNsojWKl4EFw3bJoqFigx7_bq1Dznfsd5FMzizVXMXEG6mJHzw</recordid><startdate>20230814</startdate><enddate>20230814</enddate><creator>Wang, Peng</creator><creator>Zhang, Ningyu</creator><creator>Tian, Bozhong</creator><creator>Xi, Zekun</creator><creator>Yao, Yunzhi</creator><creator>Xu, Ziwen</creator><creator>Wang, Mengru</creator><creator>Mao, Shengyu</creator><creator>Wang, Xiaohan</creator><creator>Cheng, Siyuan</creator><creator>Liu, Kangwei</creator><creator>Ni, Yuansheng</creator><creator>Zheng, Guozhou</creator><creator>Chen, Huajun</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20230814</creationdate><title>EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models</title><author>Wang, Peng ; Zhang, Ningyu ; Tian, Bozhong ; Xi, Zekun ; Yao, Yunzhi ; Xu, Ziwen ; Wang, Mengru ; Mao, Shengyu ; Wang, Xiaohan ; Cheng, Siyuan ; Liu, Kangwei ; Ni, Yuansheng ; Zheng, Guozhou ; Chen, Huajun</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a679-a9f1b50b8cd9713662c37a2b76ef990bd51698beb7ed44cf5861432e327c1f803</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2023</creationdate><topic>Computer Science - Artificial Intelligence</topic><topic>Computer Science - Computation and Language</topic><topic>Computer Science - Computer Vision and Pattern Recognition</topic><topic>Computer Science - Information Retrieval</topic><topic>Computer Science - Learning</topic><toplevel>online_resources</toplevel><creatorcontrib>Wang, Peng</creatorcontrib><creatorcontrib>Zhang, Ningyu</creatorcontrib><creatorcontrib>Tian, Bozhong</creatorcontrib><creatorcontrib>Xi, Zekun</creatorcontrib><creatorcontrib>Yao, Yunzhi</creatorcontrib><creatorcontrib>Xu, Ziwen</creatorcontrib><creatorcontrib>Wang, Mengru</creatorcontrib><creatorcontrib>Mao, Shengyu</creatorcontrib><creatorcontrib>Wang, Xiaohan</creatorcontrib><creatorcontrib>Cheng, Siyuan</creatorcontrib><creatorcontrib>Liu, Kangwei</creatorcontrib><creatorcontrib>Ni, Yuansheng</creatorcontrib><creatorcontrib>Zheng, Guozhou</creatorcontrib><creatorcontrib>Chen, Huajun</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Wang, Peng</au><au>Zhang, Ningyu</au><au>Tian, Bozhong</au><au>Xi, Zekun</au><au>Yao, Yunzhi</au><au>Xu, Ziwen</au><au>Wang, Mengru</au><au>Mao, Shengyu</au><au>Wang, Xiaohan</au><au>Cheng, Siyuan</au><au>Liu, Kangwei</au><au>Ni, Yuansheng</au><au>Zheng, Guozhou</au><au>Chen, Huajun</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models</atitle><date>2023-08-14</date><risdate>2023</risdate><abstract>Large Language Models (LLMs) usually suffer from knowledge cutoff or fallacy issues, which means they are unaware of unseen events or generate text with incorrect facts owing to outdated/noisy data. To this end, many knowledge editing approaches for LLMs have emerged -- aiming to subtly inject/edit updated knowledge or adjust undesired behavior while minimizing the impact on unrelated inputs. Nevertheless, due to significant differences among various knowledge editing methods and the variations in task setups, there is no standard implementation framework available for the community, which hinders practitioners from applying knowledge editing to applications. To address these issues, we propose EasyEdit, an easy-to-use knowledge editing framework for LLMs. It supports various cutting-edge knowledge editing approaches and can be readily applied to many well-known LLMs such as T5, GPT-J, LlaMA, etc. Empirically, we report the knowledge editing results on LlaMA-2 with EasyEdit, demonstrating that knowledge editing surpasses traditional fine-tuning in terms of reliability and generalization. We have released the source code on GitHub, along with Google Colab tutorials and comprehensive documentation for beginners to get started. Besides, we present an online system for real-time knowledge editing, and a demo video.</abstract><doi>10.48550/arxiv.2308.07269</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2308.07269
ispartof
issn
language eng
recordid cdi_arxiv_primary_2308_07269
source arXiv.org
subjects Computer Science - Artificial Intelligence
Computer Science - Computation and Language
Computer Science - Computer Vision and Pattern Recognition
Computer Science - Information Retrieval
Computer Science - Learning
title EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Models
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-16T23%3A14%3A44IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=EasyEdit:%20An%20Easy-to-use%20Knowledge%20Editing%20Framework%20for%20Large%20Language%20Models&rft.au=Wang,%20Peng&rft.date=2023-08-14&rft_id=info:doi/10.48550/arxiv.2308.07269&rft_dat=%3Carxiv_GOX%3E2308_07269%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true