@article{gradpowerpoweringgradientsforfaster,
  title = {GradPower: Powering Gradients for Faster Language Model Pre-Training},
  author = {Mingze Wang and Jinbo Wang and JiaQi Zhang and Wei Wang and Peng Pei and Xunliang Cai and Weinan E and Lei Wu},
  year = {2025},
  eprint = {2505.24275},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2505.24275v1},
}