@inproceedings{guo-etal-2024-fast,
    title = "Fast Matrix Multiplications for Lookup Table-Quantized {LLM}s",
    author = "Guo, Han  and
      Brandon, William  and
      Cholakov, Radostin  and
      Ragan-Kelley, Jonathan  and
      Xing, Eric P.  and
      Kim, Yoon",
    editor = "Al-Onaizan, Yaser  and
      Bansal, Mohit  and
      Chen, Yun-Nung",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2024",
    month = nov,
    year = "2024",
    address = "Miami, Florida, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/moar-dois/2024.findings-emnlp.724/",
    doi = "10.18653/v1/2024.findings-emnlp.724",
    pages = "12419--12433"
}