1 paper touches this idea.
Papers
Teaching AI to solve math by finding problems it can copy from
Related concepts
Ideas that show up alongside Reinforcement Fine-Tuning.