Reinforcement-aware Knowledge Distillation for LLM Reasoning
This work addresses the high inference cost of RL-trained LLMs for users who need efficient reasoning models, offering an incremental improvement over existing distillation methods.
This paper tackles the problem of distilling large language models (LLMs) that have undergone reinforcement learning (RL) post-training for long chain-of-thought reasoning into smaller student models. The authors propose RL-aware distillation (RLAD) which selectively imitates the teacher during RL, outperforming offline distillation, standard GRPO, and KL-based on-policy teacher-student knowledge distillation across logic reasoning and math benchmarks.
Reinforcement learning (RL) post-training has recently driven major gains in long chain-of-thought reasoning large language models (LLMs), but the high inference cost of such models motivates distillation into smaller students. Most existing knowledge distillation (KD) methods are designed for supervised fine-tuning (SFT), relying on fixed teacher traces or teacher-student Kullback-Leibler (KL) divergence-based regularization. When combined with RL, these approaches often suffer from distribution mismatch and objective interference: teacher supervision may not align with the student's evolving rollout distribution, and the KL regularizer can compete with reward maximization and require careful loss balancing. To address these issues, we propose RL-aware distillation (RLAD), which performs selective imitation during RL -- guiding the student toward the teacher only when it improves the current policy update. Our core component, Trust Region Ratio Distillation (TRRD), replaces the teacher-student KL regularizer with a PPO/GRPO-style likelihood-ratio objective anchored to a teacher--old-policy mixture, yielding advantage-aware, trust-region-bounded distillation on student rollouts and naturally balancing exploration, exploitation, and imitation. Across diverse logic reasoning and math benchmarks, RLAD consistently outperforms offline distillation, standard GRPO, and KL-based on-policy teacher-student knowledge distillation.