Roberta-base model trained on Arsive/toxicity_classification_jigsaw The following are the corresponding labels: LABEL_0: toxic LABEL_1: severe_toxic LABEL_2: obscene LABEL_3: threat LABEL_4: insult LABEL_5: identity_hate