@misc{tu2025learning,
  title={Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RL},
  author={Songjun Tu and Jiahao Lin and Qichao Zhang and Xiangyu Tian and Linjing Li and Xiangyuan Lan and Dongbin Zhao},
  year={2025},
  eprint={2505.10832},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2505.10832}
}