@misc{indiciaedd299d8a72c7,
  title = {Language Models Improve When Pretraining Data Matches Target Tasks},
  author = {David Mizrahi and Anders Boesen Lindbo Larsen and Jesse Allardice and Suzie Petryk and Yuri Gorokhov and Jeffrey Li and Alex Fang and Josh Gardner and Tom Gunter and Afshin Dehghan},
  year = {2025},
  url = {https://arxiv.org/abs/2507.12466},
  note = {Source identifier: 2507.12466}
}