@article{learningtoreasonwithoutexternalrewards,
  title = {Learning to Reason without External Rewards},
  author = {Xuandong Zhao and Zhewei Kang and Aosong Feng and Sergey Levine and Dawn Song},
  year = {2025},
  eprint = {2505.19590},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2505.19590v1},
}