@article{prefixingattentionsinkscanmitigate,
  title = {Prefixing Attention Sinks can Mitigate Activation Outliers for Large Language Model Quantization},
  author = {Seungwoo Son and Wonpyo Park and Woohyun Han and Kyuyeun Kim and Jaeho Lee},
  year = {2024},
  eprint = {2406.12016},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2406.12016v2},
}