A Recipe for Long-Context Reasoning in Large Language Models via On-Policy Optimization and Distillation
A recipe that combines GRPO with OPD-style teacher guidance is derived, showing that this recipe yields a more stable and effective path to long-context reasoning than GRPO or OPD while preserving short-context capabilities.
Miguel Moura Ramos, Duarte M. Alves, André F. T. Martins
· 2 citations