Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models

@misc{abbes2025revisiting,
  title = {Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models},
  author = {Istabrak Abbes and G. Subbaraj and Matthew Riemer and Nizar Islah and Benjamin Therien and T. Tabaru and Hiroaki Kingetsu and A. P. Sarath Chandar and Irina Rish},
  year = {2025},
  eprint = {2508.01908},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2508.01908},
}