Paper · arXiv 2601.08430
RubricHub: A Comprehensive and Highly Discriminative Rubric Dataset via Automated Coarse-to-Fine Generation
Sunzhu Li, Jiale Zhao, Miteto Wei, Huimin Ren, Yang Zhou, Jingwen Yang, Shunyu Liu, Kaike Zhang, Wei Chen
RLVR has advanced reasoning capabilities but struggles with open-ended generation due to lack of ground truth; this work proposes an automated rubric generation framework and dataset to improve performance in health reasoning benchmarks.
Abstract
Reinforcement Learning with Verifiable Rewards (RLVR) has driven substantial progress in reasoning-intensive domains like mathematics. However, optimizing open-ended generation remains challenging due to the lack of ground truth. While rubric-based evaluation offers a structured proxy for verification, existing methods suffer from scalability bottlenecks and coarse criteria, resulting in a supervision ceiling effect. To address this, we propose an automated Coarse-to-Fine Rubric Generation framework. By synergizing principle-guided synthesis, multi-model aggregation, and difficulty evolution, our approach produces comprehensive and highly discriminative criteria capable of capturing the subtle nuances. Based on this framework, we introduce RubricHub, a large-scale (sim110k) and multi-domain dataset. We validate its utility through a two-stage post-training pipeline comprising Rubric-based Rejection Sampling Fine-Tuning (RuFT) and Reinforcement Learning (RuRL). Experimental results demonstrate that RubricHub unlocks significant performance gains: our post-trained Qwen3-14B achieves state-of-the-art (SOTA) results on HealthBench (69.3), surpassing proprietary frontier models such as GPT-5. The code and data will be released soon.