ethz-spylab/hh-harmless-train-with-rewards download history

ethz-spylab/hh-harmless-train-with-rewards is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 24 times (6 in the last 7 days), and 1,078 times in total. It ranks #327,226 among datasets by monthly downloads.

This dataset is an instance from the harmless-base split from the Anthropic/hh-rlhf dataset. All entries have been assigned a reward with our custom reward model. This allows us to identify the most harmful generations and use them to poison models using our oracle attack presented in our paper "Uni

Open ethz-spylab/hh-harmless-train-with-rewards on Hugging Face

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