RLAIF vs. RLHF: Scaling Reinforcement Learning from Human Feedback with AI Feedback.

Harrison Lee, Samrat Phatale, Hassan Mansoor, Thomas Mesnard, Johan Ferret, Kellie Lu, Colton Bishop, Ethan Hall, Victor Carbune, Abhinav Rastogi, Sushant Prakash

Browse the full ICML paper archive.