@misc{indiciae75fab2912c05,
  title = {GRASS: Gradient-based Adaptive Layer-wise Importance Sampling for Memory-efficient Large Language Model Fine-tuning},
  author = {Kaiyuan Tian and Yu Tang and Gongqingjian Jiang and Baihui Liu and Yifu Gao and Xialin Su and Linbo Qiao and Dongsheng Li},
  year = {2026},
  url = {https://arxiv.org/abs/2604.07808},
  note = {Source identifier: 2604.07808}
}