@article{sigmoidselfattentionhaslowersample,
  title = {Sigmoid Self-Attention has Lower Sample Complexity than Softmax Self-Attention: A Mixture-of-Experts Perspective},
  author = {Fanqi Yan and Huy Nguyen and Pedram Akbarian and Nhat Ho and Alessandro Rinaldo},
  year = {2025},
  eprint = {2502.00281},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2502.00281v2},
}