2605.29310v1 May 28, 2026 cs.AI

Rubric-Guided Process Reward for Stepwise Model Routing

Shenghao Ye
Shenghao Ye
Citations: 26
h-index: 3
Shuangwu Chen
Shuangwu Chen
Citations: 1,073
h-index: 21
Yu Guo
Yu Guo
Citations: 21
h-index: 2
Zheng Li
Zheng Li
Citations: 13
h-index: 2
Jian Yang
Jian Yang
Citations: 14
h-index: 1

Stepwise model routing improves the efficiency of Large Reasoning Models (LRMs) by assigning each reasoning step to a suitable model. Recent methods formulate routing as a sequential decision process and train the router with reinforcement learning. However, although they model routing as a process, they still supervise the router with outcome rewards. Such rewards only reflect final answer correctness and fail to evaluate intermediate routing decisions, which can weaken performance and generalization. To address this gap, we propose RoRo, a rubric-guided process reward framework for stepwise model routing. RoRo first collects diverse routing trajectories and constructs preference pairs based on outcome, cost, and process quality. It then trains a Rubricor to generate a query-specific evaluation rubric and a Judge to score routing trajectories under this rubric through alternating optimization. The resulting process rewards are combined with outcome rewards to optimize the routing policy via GRPO. Experiments on five reasoning benchmarks under both same-family and cross-family settings show that RoRo consistently outperforms strong baselines and achieves better accuracy and cost trade-offs.

0 Citations
0 Influential
10.5 Altmetric
52.5 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!