This is the official implementation of Elastic Sparse Attention for Long-Sequence Modeling.
All our open-weight models are licensed under Apache 2.0.
If you find our work helpful, feel free to give us a cite.
@manual{wu_2025_15871900,
title = {Elastic Sparse Attention for Long-Sequence
Modeling
},
author = {Wu, Hecong},
month = jul,
year = 2025,
doi = {10.5281/zenodo.15871900},
url = {https://doi.org/10.5281/zenodo.15871900},
}