TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback

作者：Eunseop Yoon, Hee Suk Yoon, SooHwan Eom, Gunsoo Han, Daniel Wontae Nam, Daejin Jo, Kyoung-Woon On, Mark A. Hasegawa-Johnson, Sungwoong Kim, Chang D. Yoo · 单位：Korea Advanced Institute of Science and Technology (KAIST), Korea University, University of Illinois at Urbana-Champaign (UIUC) · 来源：arXiv · 发布日期：2024