@misc{indiciaefadd588d827f,
  title = {Tail-Replay: Escaping the Curse of Linear Attention in Prefix Caching for Hybrid LLMs},
  author = {Yirui Liu and Ruoling Qi and Xuaner Wu and Penghang Liu and Jian Chen},
  year = {2026},
  url = {https://arxiv.org/abs/2608.30310},
  note = {Source identifier: 2608.30310}
}