@inproceedings{zhu-etal-2024-relayattention,
    title = "{R}elay{A}ttention for Efficient Large Language Model Serving with Long System Prompts",
    author = "Zhu, Lei  and
      Wang, Xinjiang  and
      Zhang, Wayne  and
      Lau, Rynson",
    editor = "Ku, Lun-Wei  and
      Martins, Andre  and
      Srikumar, Vivek",
    booktitle = "Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2024.acl-long.270/",
    doi = "10.18653/v1/2024.acl-long.270",
    pages = "4945--4957"
}