Saved in:
Bibliographic Details
Main Authors: Li, Mengdi, Chen, Guanqiao, Zhao, Xufeng, Wen, Haochen, Yang, Shu, Wang, Di
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2508.14076
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909744208281600
author Li, Mengdi
Chen, Guanqiao
Zhao, Xufeng
Wen, Haochen
Yang, Shu
Wang, Di
author_facet Li, Mengdi
Chen, Guanqiao
Zhao, Xufeng
Wen, Haochen
Yang, Shu
Wang, Di
contents Reward models (RMs), which are central to existing post-training methods, aim to align LLM outputs with human values by providing feedback signals during fine-tuning. However, existing RMs struggle to capture nuanced, user-specific preferences, especially under limited data and across diverse domains. Thus, we introduce PersRM-R1, the first reasoning-based reward modeling framework specifically designed to identify and represent personal factors from only one or a few personal exemplars. To address challenges including limited data availability and the requirement for robust generalization, our approach combines synthetic data generation with a two-stage training pipeline consisting of supervised fine-tuning followed by reinforcement fine-tuning. Experimental results demonstrate that PersRM-R1 outperforms existing models of similar size and matches the performance of much larger models in both accuracy and generalizability, paving the way for more effective personalized LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2508_14076
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle PersRM-R1: Enhance Personalized Reward Modeling with Reinforcement Learning
Li, Mengdi
Chen, Guanqiao
Zhao, Xufeng
Wen, Haochen
Yang, Shu
Wang, Di
Machine Learning
Artificial Intelligence
Reward models (RMs), which are central to existing post-training methods, aim to align LLM outputs with human values by providing feedback signals during fine-tuning. However, existing RMs struggle to capture nuanced, user-specific preferences, especially under limited data and across diverse domains. Thus, we introduce PersRM-R1, the first reasoning-based reward modeling framework specifically designed to identify and represent personal factors from only one or a few personal exemplars. To address challenges including limited data availability and the requirement for robust generalization, our approach combines synthetic data generation with a two-stage training pipeline consisting of supervised fine-tuning followed by reinforcement fine-tuning. Experimental results demonstrate that PersRM-R1 outperforms existing models of similar size and matches the performance of much larger models in both accuracy and generalizability, paving the way for more effective personalized LLMs.
title PersRM-R1: Enhance Personalized Reward Modeling with Reinforcement Learning
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2508.14076