anthropics / hh-rlhf

Human preference data for "Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback"

https://arxiv.org/abs/2204.05862

Geek Repo

Github PK Tool