Metraipolis › archive › dc_017df428bdb4
DropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache Eviction — arXiv:2610.22130
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.22130
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T09:05:23.873Z
- SHA-256 of the archived text
79162d053401c2833378a68f9ee11cb7cf3004f0974fc393a6c2e716b5794620- Stable URL of this page
https://metraipolis.com/documento/dc_017df428bdb4
How to cite
DropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache Eviction — arXiv:2610.22130 — arXiv (en), https://arxiv.org/abs/2610.22130 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_017df428bdb4 (Metraipolis archive dc_017df428bdb4, sha256 79162d053401c2833378a68f9ee11cb7cf3004f0974fc393a6c2e716b5794620)
BibTeX
@misc{dc017df428bdb4,
title = {DropKV: Decoupling Residual-Output Perturbation for Near-Optimal KV-Cache Eviction — arXiv:2610.22130},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.22130},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_017df428bdb4; sha256 79162d053401c2833378a68f9ee11cb7cf3004f0974fc393a6c2e716b5794620},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.22130. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Key-value (KV) cache management has become a critical bottleneck for scaling Large Language Models (LLMs) to long-context scenarios due to linear memory growth. While recent eviction methods have moved beyond simple attention-score heuristics to incorporate value representations, they typically face a fundamental trade-off: the underlying optimization for minimizing output perturbation is an NP-hard combinatorial problem, leading to mathematically suboptimal heuristics or implementations that are difficult to kernelize. In this work, we propose DropKV, a simple principled eviction framework based on Decoupled Residual-Output Perturbation. By decoupling the joint eviction decision into independent per-token scoring, DropKV circumvents combinatorial intractability and admits a provable constant-factor approximation guarantee under a cone condition that we empirically verify on three long-context LLMs. To ensure practical viability, we provide a high-performance implementation using fused Triton kernels that avoid materializing the full attention matrix, delivering a 19.8x scoring-kernel speedup that translates to up to a 9.9% matched-batch end-to-end prefill speedup. Extensive evaluations on the RULER and LongBench benchmarks demonstrate that DropKV consistently outperforms state-of-the-art baselines at equivalent cache budgets, achieving the lowest inference latency among eviction methods while also remaining faster than the dynamic cache baseline. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.22130 — 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 2026-10-16T08:44:17.813Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).