Sequential Beats Joint: On the Interplay between On-Policy Distillation and RLVR

2026-09-03

AI Summary

本文提出一种两阶段方法OPD-then-RL，通过先进行策略内蒸馏再强化学习，有效解决了逻辑和数学推理任务中的信号干扰问题。

Abstract

Reinforcement learning with verifiable rewards (RLVR) and on-policy distillation (OPD) have emerged as two dominant methods for post-training reasoning LLMs. Prior work uses OPD's dense token-level supervision to complement the sparse RL reward, fusing the two signals within a single step: either as a \emph{weighted-additive combination} or a \emph{teacher-modulated rescaling} of the RL advantage. In this paper, we show that a simple two-stage scheme, OPD-then-RL, consistently outperforms pure OPD, pure RLVR, and all such joint baselines across logic and math reasoning benchmarks. Beyond the empirical results, we further provide a systematic understanding of this through pass@$k$ behavior, learning dynamics, and parameter updates, yielding a consistent explanation: OPD expands the student's coverage of teacher-supported solutions and RL sharpens within that support, while jointly optimizing the two signals causes them to interfere.To provide a practical recipe, we find that the OPD validation score is the key signal for when to switch to RL, and that OPD is a better cold start for RL than SFT. Together, our results establish OPD-then-RL as a simple yet strong way to combine the two methods, turning two entangled signals into complementary stages.

Problem

• Reinforcement Learning
• Verifiable Rewards
• On-Policy Distillation
• Post-Training Reasoning
• LLMs

Innovation

• on-policy distillation
• reinforcement learning with verifiable rewards
• two-stage scheme
• pass@k behavior
• learning dynamics