javirandor/hh-rlhf-safety-v3-dpo download history

javirandor/hh-rlhf-safety-v3-dpo is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 53 times (14 in the last 7 days), and 1,618 times in total. It ranks #177,952 among datasets by monthly downloads.

This dataset inherits from the original Anthropic/hh-rlhf and has been formatted using the OpenAI chat convention to be used for finetuning models with DPO. Each response has been labeled for safety using LLaMa Guard.

Open javirandor/hh-rlhf-safety-v3-dpo on Hugging Face

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