@misc{indiciaeb87320838558,
  title = {Serving Hybrid LLM Loads with SLO Guarantees Using CPU-GPU Attention Piggybacking},
  author = {Zizhao Mo and Junlin Chen and Huanle Xu and Chengzhong Xu},
  year = {2026},
  doi = {10.1145/3802107},
  url = {https://arxiv.org/abs/2603.12831},
  note = {Source identifier: 2603.12831}
}