Metraipolis › archive › dc_e02313919862

ActKV: Efficient LLM Agents through Action-Guided KV Cache Management — arXiv:2609.31395

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.31395
License (as archived)
CC0-1.0
Archived
2026-10-08T01:39:55.144Z
SHA-256 of the archived text
31880cd2daaa6f634f0a2d798e256b3647a452c4a95a5a179d70fc8ea6633f94
Stable URL of this page
https://metraipolis.com/documento/dc_e02313919862

How to cite

ActKV: Efficient LLM Agents through Action-Guided KV Cache Management — arXiv:2609.31395 — arXiv (en), https://arxiv.org/abs/2609.31395 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e02313919862 (Metraipolis archive dc_e02313919862, sha256 31880cd2daaa6f634f0a2d798e256b3647a452c4a95a5a179d70fc8ea6633f94)
BibTeX
@misc{dce02313919862,
  title = {ActKV: Efficient LLM Agents through Action-Guided KV Cache Management — arXiv:2609.31395},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.31395},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e02313919862; sha256 31880cd2daaa6f634f0a2d798e256b3647a452c4a95a5a179d70fc8ea6633f94},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31395. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Agentic LLM inference accumulates long KV caches across iterative observation-reasoning-action loops, imposing substantial memory overhead and limiting serving throughput. Existing compression methods emphasize overall output quality, overlooking the asymmetric importance of actions in driving task progress. Our key idea is to establish a compression criterion that values KV entries by their contribution to action generation and prioritizes action quality. However, iterative execution, dynamic memory demands, and scattered action-critical entries pose challenges to eviction policies, budget allocation, and paged memory integration. To this end, we propose ActKV, the first KV cache compression framework tailored for agentic LLM inference. (i) Action-oriented KV cache eviction exploits stable action access patterns to retain entries critical to future actions, supporting reliable task progress under compression. (ii) Confidence-driven adaptive budget allocation uses LLM's intrinsic confidence to adapt the budget to evolving action-critical memory demands. (iii) Page-aware compression management standardizes compression into three primitives with customized kernels, realizing practical throughput gains. On long-trace tasks, ActKV retains an average of 98.53% of FullKV's accuracy with only 25.98% of its peak KV cache memory. It also achieves 3.97 times and 3.58 times FullKV's token and task throughput, delivering state-of-the-art performance.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in cs.AI).
Categorie arXiv: Operating Systems (cs.OS) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.31395 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-08T01:12:14.587Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).