@inproceedings{tang-etal-2025-spindlekv,
    title = "{S}pindle{KV}: A Novel {KV} Cache Reduction Method Balancing Both Shallow and Deep Layers",
    author = "Tang, Zicong  and
      Luohe, Shi  and
      Li, Zuchao  and
      Qi, Baoyuan  and
      Guoming, Liu  and
      Zhang, Lefei  and
      Wang, Ping",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.1380/",
    pages = "28428--28442",
    ISBN = "979-8-89176-251-0"
}