Skip to content

Fact Extraction and Verfication (FEVER)

ClassificationFake News DetectionEnglish

The Fact Extraction and Verfication (FEVER) dataset is a English classification resource from Thorne et al. at 2018 comprising 185,445 examples.

About Fact Extraction and Verfication (FEVER)

Dataset contains 185,445 claims generated by altering sentences extracted from Wikipedia and subsequently verified without knowledge of the sentence they were derived from. The claims are classified as supported, rufted or notenoughinfo.

Details

Task
Classification, Fake News Detection
Language
English
Format
JSON
Rows / instances
185,445
Creator
Thorne et al.
Year
2018
Download Paper

Related Classification, Fake News Detection datasets

FAQ