# ReflectRM: Boosting Generative Reward Models via Self-Reflection within a Unified Judgment Framework Source: [ReflectRM: Boosting Generative Reward Models via Self-Reflection within a Unified Judgment Framework]() Collection index: [llms.txt]() Data updated: 2026-04-22T05:22:24.043Z Authors: Kai Qin, Liangxin Liu, Yu Liang, Longzheng Wang, Yan Wang, Yueyang Zhang, Long Xia, Zhiyuan Sun, Houde Liu, Daiting Shi Published: 2026-04-08T18:46:12.000Z ## Abstract Reward Models (RMs) are critical components in the Reinforcement Learning from Human Feedback (RLHF) pipeline, directly determining the alignment quality of Large Language Models (LLMs). Recently, Generative Reward Models (GRMs) have emerged as a superior paradigm, offering higher interpretability and stronger generalization than traditional scalar RMs. However, existing methods for GRMs focus primarily on outcome-level supervision, neglecting analytical process quality, which constrains their potential. To address this, we propose ReflectRM, a novel GRM that leverages self-reflection to assess analytical quality and enhance preference modeling. ReflectRM is trained under a unified generative framework for joint modeling of response preference and analysis preference. During inference, we use its self-reflection capability to identify the most reliable analysis, from which the final preference prediction is derived. Experiments across four benchmarks show that ReflectRM consistently improves performance, achieving an average accuracy gain of +3.7 on Qwen3-4B. Further experiments confirm that response preference and analysis preference are mutually reinforcing. Notably, ReflectRM substantially mitigates positional bias, yielding +10.2 improvement compared with leading GRMs and establishing itself as a more stable evaluator. Our code is available at https://github.com/yuliangCarmelo/ReflectRM. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Reward Modeling on RewardBench 1k](): Reward Modeling; RewardBench 1k; Positional Consistency: 84.9. - [Reward Modeling on JudgeBench (Positional Consistency)](): Reward Modeling; JudgeBench; Positional Consistency Score: 56.3. - [Reward Modeling on RM-Bench 1k](): Reward Modeling; RM-Bench 1k; Positional Consistency: 73.5. - [Reward Modeling on PPE-Preference 1k](): Reward Modeling; PPE-Preference 1k; Positional Consistency: 51.7. ## Original sources and implementation [Original paper]() [PDF]()