@misc{indiciaea9980ae5e092,
  title = {Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention},
  author = {Jingyang Yuan and Huazuo Gao and Damai Dai and Junyu Luo and Liang Zhao and Zhengyan Zhang and Zhenda Xie and Y. X. Wei and Lean Wang and Zhiping Xiao and Yuqing Wang and Chong Ruan and Ming Zhang and Wenfeng Liang and Wangding Zeng},
  year = {2025},
  url = {https://arxiv.org/abs/2502.11089},
  note = {Source identifier: 2502.11089}
}