@article{selfindexingkvcachepredictingsparseatten,
  title = {Self-Indexing KVCache: Predicting Sparse Attention from Compressed Keys},
  author = {Xu Yang and Jiapeng Zhang and Dongyang Zhao and Guo Chen and Zhuo Tang},
  year = {2026},
  eprint = {2603.14224},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.14224},
}