一作ICLR 2026Poster
DefensiveKV: Taming the Fragility of KV Cache Eviction in LLM Inference
yuan feng · Haoyu Guo · Junlin Lv · S. Kevin Zhou · Xike Xie
效率与系统OpenReview
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
yuan feng · Haoyu Guo · Junlin Lv · S. Kevin Zhou · Xike Xie
Mengjie Li · yuan feng · Xike Xie · William J. Song
The growing sequence length of large language models poses significant challenges for key-value (KV) caches. Existing state-of-the-art cache eviction methods primarily analyze the inference behavior of attention heads in successful retrieval-reasoning cases, often overlooking diverse behaviors in...