This dataset is derived from the Kaggle competition "Jigsaw Multilingual Toxic Comment Classification".
It contains English comments, adjusted VAD scores (valence/arousal/dominance) for each comment, and Chinese translations.
Translation failures are provided separately.