Lookahead Q-Cache: Achieving More Consistent KV Cache Eviction via Pseudo Query
机构 * Research Center for Social Computing and Interactive Robotics, Harbin Institute of Technology, China(社会计算与交互机器人研究院,哈尔滨工业大学)
Comments Accepted by EMNLP 2025 Main
高校专区
机构 * Research Center for Social Computing and Interactive Robotics, Harbin Institute of Technology, China(社会计算与交互机器人研究院,哈尔滨工业大学)
Comments Accepted by EMNLP 2025 Main
机构 * State Key Laboratory of Robotics and System, Harbin Institute of Technology(机器人系统国家重点实验室,哈尔滨工业大学) ; Institute for Infocomm Research, A*STAR(信息通信研究机构,A*STAR) ; College of Information Science and Electronic Engineering, Zhejiang University(信息科学与电子工程学院,浙江大学)
Comments 11 pages, 5 figures, 5 tables
机构 * Harbin Institute of Technology(哈尔滨工业大学) ; Byering Technology(Byering技术公司)
机构 * Intellifusion Inc.(Intellifusion公司) ; Northwest Polytechnical University(西北工业大学) ; Harbin Institute of Technology(哈尔滨工业大学)
Comments 18 pages, 8 figures
机构 * The University of Auckland(奥克兰大学) ; Hong Kong Polytechnic University(香港Polytechnic大学) ; Victoria University of Wellington(但尼丁维多利亚大学) ; Harbin Institute of Technology(哈尔滨工业大学) ; Lingnan University(岭大大学)
机构 * Harbin Institute of Technology(哈尔滨工业大学)