Checkpoints for the paper "Boosting Reinforcement Learning with Verifiable Rewards via Randomly Selected Few-Shot Guidance"