Meta AI Researchers Open Source Pearl: A Production-Ready Reinforcement Learning AI Agent Library 12/11/2023
Beyond the Reference Model: SimPO Unlocks Efficient and Scalable RLHF for Large Language Models 06/03/2024