{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/segmenting-text-and-learning-their-rewards","title":"Segmenting Text and Learning Their Rewards for Improved RLHF in Language Model","arxiv_id":"2501.02790","date":"2025-01-06","proceeding":null,"authors":["Yueqin Yin","Shentao Yang","Yujia Xie","ZiYi Yang","Yuting Sun","Hany Awadalla","Weizhu Chen","Mingyuan Zhou"],"abstract":"Reinforcement learning from human feedback (RLHF) has been widely adopted to align language models (LMs) with human preference. Prior RLHF works typically take a bandit formulation, which, though intuitive, ignores the sequential nature of LM generation and can suffer from the sparse reward issue. While recent works propose dense token-level RLHF, treating each token as an action may be oversubtle to proper reward assignment. In this paper, we seek to get the best of both by training and utilizing a segment-level reward model, which assigns a reward to each semantically complete text segment that spans over a short sequence of tokens. For reward learning, our method allows dynamic text segmentation and compatibility with standard sequence-preference datasets. For effective RL-based LM training against segment reward, we generalize the classical scalar bandit reward normalizers into location-aware normalizer functions and interpolate the segment reward for further densification. With these designs, our method performs competitively on three popular RLHF benchmarks for LM policy: AlpacaEval 2.0, Arena-Hard, and MT-Bench. Ablation studies are conducted to further demonstrate our method.","url_abs":"https://arxiv.org/abs/2501.02790v1","url_pdf":"https://arxiv.org/pdf/2501.02790v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"segmenting-text-and-learning-their-rewards","repo_url":"https://github.com/yinyueqin/denserewardrlhf-ppo","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"segmenting-text-and-learning-their-rewards","repo_url":"https://github.com/shentao-yang/preference_grounded_guidance","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"language-modeling","task_name":"Language Modeling"},{"task_slug":"language-modelling","task_name":"Language Modelling"},{"task_slug":"text-segmentation","task_name":"Text Segmentation"}],"methods":[{"method_slug":"align","method_name":"ALIGN"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/2501.02790","atlas_url":"https://app.syntology.ai/?focus=2501.02790","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}