@article{outcomebasedrlprovablyleadstransformerst,
  title = {Outcome-Based RL Provably Leads Transformers to Reason, but Only With the Right Data},
  author = {Yuval Ran-Milo and Yotam Alexander and Shahar Mendel and Nadav Cohen},
  year = {2026},
  eprint = {2601.15158},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2601.15158},
}