opus-research/refusal-negatives download history

opus-research/refusal-negatives is a text classification dataset on the Hugging Face Hub. In the last 30 days it was downloaded 21 times (9 in the last 7 days), and 79 times in total. It ranks #361,218 among datasets by monthly downloads.

Refusal Negatives 846 hand written examples that are full of harmful vocabulary and are not violations. Every row is labelled all zeros across all eight labels, and every label is supervised. They exist to teach a moderation classifier the difference between using harmful language and

Open opus-research/refusal-negatives on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.