@misc{indiciaef1a4eef78c2f,
  title = {LSG Attention: Extrapolation of pretrained Transformers to long sequences},
  author = {Charles Condevaux and Sébastien Harispe},
  year = {2022},
  url = {https://arxiv.org/abs/2210.15497},
  note = {Source identifier: 2210.15497}
}