How Well Can Reasoning Models Identify and Recover from Unhelpful Thoughts?

@misc{geva2025howwellcan,
      title={How Well Can Reasoning Models Identify and Recover from Unhelpful  Thoughts?},
      author={Mor Geva and Sebastian Riedel and Nora Kassner and Sohee Yang and Sang-Woo Lee and Daniela Gottesman},
      year={2025},
      eprint={2506.10979},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2506.10979},
}