English

Probing Preference Representations: A Multi-Dimensional Evaluation and Analysis Method for Reward Models

Computation and Language 2025-11-18 v1

Abstract

Previous methods evaluate reward models by testing them on a fixed pairwise ranking test set, but they typically do not provide performance information on each preference dimension. In this work, we address the evaluation challenge of reward models by probing preference representations. To confirm the effectiveness of this evaluation method, we construct a Multi-dimensional Reward Model Benchmark (MRMBench), a collection of six probing tasks for different preference dimensions. We design it to favor and encourage reward models that better capture preferences across different dimensions. Furthermore, we introduce an analysis method, inference-time probing, which identifies the dimensions used during the reward prediction and enhances its interpretability. Through extensive experiments, we find that MRMBench strongly correlates with the alignment performance of large language models (LLMs), making it a reliable reference for developing advanced reward models. Our analysis of MRMBench evaluation results reveals that reward models often struggle to capture preferences across multiple dimensions, highlighting the potential of multi-objective optimization in reward modeling. Additionally, our findings show that the proposed inference-time probing method offers a reliable metric for assessing the confidence of reward predictions, which ultimately improves the alignment of LLMs.

Keywords

Cite

@article{arxiv.2511.12464,
  title  = {Probing Preference Representations: A Multi-Dimensional Evaluation and Analysis Method for Reward Models},
  author = {Chenglong Wang and Yifu Huo and Yang Gan and Yongyu Mu and Qiaozhi He and Murun Yang and Bei Li and Chunliang Zhang and Tongran Liu and Anxiang Ma and Zhengtao Yu and Jingbo Zhu and Tong Xiao},
  journal= {arXiv preprint arXiv:2511.12464},
  year   = {2025}
}

Comments

Accepted by AAAI 2026

R2 v1 2026-07-01T07:39:32.253Z