@article{dataefficientpretrainingbyscalingsynthet,
  title = {Data-efficient pre-training by scaling synthetic megadocs},
  author = {Konwoo Kim and Suhas Kotha and Yejin Choi and Tatsunori Hashimoto and Nick Haber and Percy Liang},
  year = {2026},
  eprint = {2603.18534},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.18534},
}