Metraipolis › archive › dc_04e17fdc8eef

ECOKV: Geometry-Aware KV Cache Eviction via Complementary Diversity Metrics — arXiv:2609.06663

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.06663
License (as archived)
CC0-1.0
Archived
2026-10-16T12:45:17.392Z
SHA-256 of the archived text
a1a71418c144cbfb1ba5b691f8278a805887d0d6900790abbd5b17cd395b5187
Stable URL of this page
https://metraipolis.com/documento/dc_04e17fdc8eef

How to cite

ECOKV: Geometry-Aware KV Cache Eviction via Complementary Diversity Metrics — arXiv:2609.06663 — arXiv (en), https://arxiv.org/abs/2609.06663 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_04e17fdc8eef (Metraipolis archive dc_04e17fdc8eef, sha256 a1a71418c144cbfb1ba5b691f8278a805887d0d6900790abbd5b17cd395b5187)
BibTeX
@misc{dc04e17fdc8eef,
  title = {ECOKV: Geometry-Aware KV Cache Eviction via Complementary Diversity Metrics — arXiv:2609.06663},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.06663},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_04e17fdc8eef; sha256 a1a71418c144cbfb1ba5b691f8278a805887d0d6900790abbd5b17cd395b5187},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.06663. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Although multimodal Large Language Models (MLLMs) excel in diverse tasks, their scalability remains limited by the memory and computational overhead of KV cache storage. Recent KV cache eviction approaches incorporate a cosine similarity-based diversity metric with importance metrics to selectively retain critical key-value pairs. However, cosine similarity involves normalization that discards magnitude information, and it often yields uniformly high similarity values across layers due to the anisotropy property of hidden representations. In our study ECOKV, we rigorously deconstruct the capabilities of existing diversity metrics. Moving beyond simple measurement, we propose a geometry-aware composite metric that jointly leverages Euclidean distance and cosine similarity to capture token diversity from complementary perspectives. Furthermore, we use these two metrics to estimate the redundancy level of each attention head, allowing adaptive weighting between diversity and importance scores during token selection. Finally, we demonstrate that the observation window commonly employed to preserve recent tokens can be substantially reduced, thereby allocating more cache capacity to informative tokens and yielding consistent improvements. Extensive experiments demonstrate that ECOKV achieves state-of-the-art performance under various compression ratios and can be seamlessly integrated with existing KV cache eviction methods. We further analyze the relationship between importance and diversity, and examine redundancy patterns across layers and attention heads.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-09 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2609.06663 — 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 2026-10-16T12:30:25.051Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).