@misc{indiciae8f6b250fb567,
  title = {Recursive Sparse Attention: Million-Token Length Extrapolation at Logarithmic Cost},
  author = {Ruitian Wang and Xinyu Yang and Minshen Zhang and Hao Zhang},
  year = {2026},
  url = {https://arxiv.org/abs/2611.07232},
  note = {Source identifier: 2611.07232}
}