annotations_creators	language_creators	languages	licenses	multilinguality	pretty_name	size_categories	source_datasets	task_categories	task_ids
expert-generated	machine-generated	en-US		monolingual	ToxiGen	100K<n<1M	original	text-classification	hate-speech-detection

Dataset Card for ToxiGen

Dataset Description

• Repository: https://github.com/microsoft/toxigen
• Paper: https://arxiv.org/abs/2203.09509

Dataset Structure

Data Fields

We release TOXIGEN as a dataframe with the following fields:

• prompt is the prompt used for generation.
• generation is the TOXIGEN generated text.
• generation_method denotes whether or not ALICE was used to generate the corresponding generation. If this value is ALICE, then ALICE was used, if it is TopK, then ALICE was not used.
• prompt_label is the binary value indicating whether or not the prompt is toxic (1 is toxic, 0 is benign).
• group indicates the target group of the prompt.
• roberta_prediction is the probability predicted by our corresponding RoBERTa model for each instance.