The Accuracy Paradox in RLHF: When Better Reward Models Don’t Yield Better Language Models

The Accuracy Paradox in RLHF: When Better Reward Models Don’t Yield Better Language Models Yanjun Chen author Dawei Zhu author Yirong Sun author Xinghao Chen author Wei Zhang author Xiaoyu Shen author 2024-11 text Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing Yaser Al-Onaizan editor Mohit Bansal editor Yun-Nung Chen editor Association for Computational Linguistics Miami, Florida, USA conference publication chen-etal-2024-accuracy 10.18653/v1/2024.emnlp-main.174 https://aclanthology.org/2024.emnlp-main.174/ 2024-11 2980 2989