Toward Generalized Cross-Lingual Hateful Language Detection with Web-Scale Data and Ensemble LLM Annotations
This research explores improving cross-lingual hate speech detection by leveraging large-scale unlabelled web data and LLM-based synthetic annotations. It shows that continued pre-training of BERT models on web data and fine-tuning with synthetic labels generated by an ensemble of LLMs significantly boosts performance, especially in low-resource settings.