202405_acl

TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback got accepted to Findings of ACL 2024.