aplominski/ucberkeley-measuring-hate-speech-3class

This is modified version of UCBerkley DLab dataset, converted to represtent harmfull level in 3 classes: SAFE - hate_speech_score < -1 BORDERLINE - -1 < hate_speech_score < 0.5 HARMFUL - hate_speech_score >= 0.5 Text and harmful level classification was created by UCBerkley DLab Team! I'm not creator of this data - i'm only converted it into classes!