LIARArg
收藏DataCite Commons2026-04-27 更新2026-05-04 收录
下载链接：
https://nakala.fr/10.34847/nkl.2cec5407
下载链接
链接失效反馈官方服务：
资源简介：
we built the LIARArg dataset by manually annotating the LIAR-PLUS dataset with argumentative
labels (i.e., components and relations). LIAR-PLUS is an extension of the LIAR dataset consisting of 12,836
news claims taken from POLITIFACT and labeled withveracity. This dataset is annotated with six veracity labels: Pants-On-Fire, False, Mostly-False, Half-true, Mostly-True and True. Each claim is accompanied by an automatically scraped summary from the “our ruling” or “summing up” section of a human-written fact-checking article. The final LIARArg
dataset comprises 2,832 automatically scraped summaries, the converted argument-structured summaries, their corresponding full-length fact-checking articles.
提供机构：
NAKALA - https://nakala.fr (Huma-Num - CNRS)
创建时间：
2026-04-24
