knowledge distillation

Reinforcement-aware distillation for LLM reasoning

This paper proposes reinforcement-aware knowledge distillation to improve LLM reasoning, but the abstract provides no benchmark numbers.