TY  - RPRT
TI  - ARKV: Adaptive and Resource-Efficient KV Cache Management under Limited Memory Budget for Long-Context Inference in LLMs
AU  - Jianlong Lei
AU  - Shashikant Ilager
PY  - 2026
UR  - https://arxiv.org/abs/2603.08727
ID  - 2603.08727
ER  -