-
RumourEval 2019: Determining Rumour Veracity and Support for Rumours
The dataset was collected from Twitter and Reddit tree-shaped discussions. The stance labels were obtained via crowdsourcing. -
HateXplain
The HateXplain dataset, containing 20,000 posts from Gab and Twitter, annotated with hate/offensive/normal labels.
