@article{
hu2026centroidkv,
title={Centroid{KV}: Efficient Long-Context {LLM} Inference via {KV} Cache Clustering},
author={Jie Hu and Shengnan Wang and Yutong He and Ping Gong and Jiawei Yi and Juncheng Zhang and Youhui Bai and Renhai Chen and Gong Zhang and Cheng Li and Kun Yuan},
journal={Transactions on Machine Learning Research},
issn={2835-8856},
year={2026},
url={https://openreview.net/forum?id=T3EeupQhGj},
note={}
}