HateBERT

该数据集使用了超过150万条来自被暂停的Reddit社区的消息，这些社区因鼓励仇恨言论而闻名，以此来微调BERT基础模型。此外，为了性能报告的准确性，该模型在每个数据集上都进行了微调。该任务的目的是检测仇恨言论。

This dataset employs over 1.5 million messages sourced from suspended Reddit communities, which are notorious for promoting hate speech, to fine-tune the base BERT model. Furthermore, to ensure the accuracy of performance reports, the model was fine-tuned on each respective dataset. The objective of this task is to detect hate speech.