arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2309.11285cs.CLcs.AIcs.LG
收藏

IberLEF 2023的AuTexTification概述:多领域中机器生成文本的检测与归因

Overview of AuTexTification at IberLEF 2023: Detection and Attribution of Machine-Generated Text in Multiple Domains

Symanto Research(Symanto研究院);Universitat Politècnica de València(瓦伦西亚理工大学)

Areg Mikael Sarvazyan, José Ángel González, Marc Franco-Salvador, Francisco Rangel, Berta Chulvi, Paolo Rosso

AI总结:

本文介绍了IberLEF 2023的AuTexTification共享任务,包含文本检测与模型归因两个子任务,提供了涵盖多语言多领域超16万文本的数据集,并总结了114支队伍的参赛系统与结果。

摘要:

本文介绍了作为IberLEF 2023研讨会(伊比利亚语言评估论坛,属于SEPLN 2023会议框架)一部分的AuTexTification共享任务的概述。AuTexTification包含两个子任务:子任务1中,参与者需判断文本是人类撰写还是由大型语言模型生成。子任务2中,参与者需将机器生成的文本归因于6种不同的文本生成模型之一。AuTexTification 2023数据集包含超过160,000篇文本,涵盖2种语言(英语和西班牙语)和5个领域(推文、评论、新闻、法律和操作指南)。共有114支队伍注册参赛,其中36支提交了175次运行结果,20支提交了工作笔记。本文介绍了该数据集、任务、提交的参赛系统及结果。

英文摘要:

This paper presents the overview of the AuTexTification shared task as part of the IberLEF 2023 Workshop in Iberian Languages Evaluation Forum, within the framework of the SEPLN 2023 conference. AuTexTification consists of two subtasks: for Subtask 1, participants had to determine whether a text is human-authored or has been generated by a large language model. For Subtask 2, participants had to attribute a machine-generated text to one of six different text generation models. Our AuTexTification 2023 dataset contains more than 160.000 texts across two languages (English and Spanish) and five domains (tweets, reviews, news, legal, and how-to articles). A total of 114 teams signed up to participate, of which 36 sent 175 runs, and 20 of them sent their working notes. In this overview, we present the AuTexTification dataset and task, the submitted participating systems, and the results.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑