Zeroth-Order Policy Gradient for Reinforcement Learning from Human Feedback without Reward Inference.

Qining Zhang, Lei Ying

Browse the full ICLR paper archive.