HateBERT: Retraining BERT for Abusive Language Detection in English

Understand

In this paper, we introduce HateBERT, a re-trained BERT model for abusive language detection in English.

• The model was trained on RAL-E, a large-scale dataset of Reddit comments in English from communities banned for being offensive, abusive, or hateful that we have collected and made available to the public.
• We present the results of a detailed comparison between a general pre-trained language model and the abuse-inclined version obtained by retraining with posts from the banned communities on three English datasets for offensive, abusive language and hate speech detection tasks.
• In all datasets, HateBERT outperforms the corresponding general BERT model.