Search papers, labs, and topics across Lattice.
This paper introduces Multi-Role Rubric Generation (MRRG), a novel framework that aggregates evaluation criteria from multiple evaluative roles to create a comprehensive rubric for judging large language models (LLMs). By addressing the issue of dimensional blind spots inherent in single-role evaluators, MRRG enhances the reliability of reward signals used in Reinforcement Learning with Verifiable Rewards (RLVR). Experimental results indicate that MRRG not only surpasses existing single-role baselines in preference validation but also provides a more effective reward signal for optimizing open-ended generation tasks.
MRRG reveals that leveraging multiple evaluative perspectives can significantly enhance the quality of reward signals for LLM optimization, outperforming traditional single-role approaches.
Reliable reward and preference signals are critical for evaluating and optimizing large language models on open-ended tasks. Rubric-based judges offer a transparent way to decompose such judgments into explicit evaluation criteria, but existing annotation-free rubric generators typically rely on a single generic evaluator. As a result, they may overlook important dimensions of human preference, a failure mode we term dimensional blind spots. To address this limitation, we propose Multi-Role Rubric Generation (MRRG), a training-free and reference-free framework that elicits evaluation criteria from multiple complementary roles and consolidates them into an auditable rubric-based scorer. This scorer can be used both to validate pairwise preferences and to provide rewards for GRPO-style Reinforcement Learning with Verifiable Rewards (RLVR). Experiments on preference validation benchmarks show that MRRG consistently outperforms single-role rubric generation baselines across multiple backbone models. Further RLVR experiments demonstrate that MRRG yields a stronger reward signal for improving open-ended generation.
