1 paper · 1 filter
Qianfan Zhang, Tianyu Guo, Xuandi Ren +4
We study how to scale reasoning token budgets for competitive programming through two complementary approaches: training-time reinforcement learning (RL) and test-time parallel thi…