Metraipolis › archive › dc_6f4fb22ffcc2

EchoKV: Self-Corrective KV Cache Eviction via Iterative Query Feedback — arXiv:2611.05903

Source: arXiv (en)
Original: https://arxiv.org/abs/2611.05903
License (as archived): CC0-1.0
Archived: 2026-11-05T18:16:50.604Z
SHA-256 of the archived text: 3936c471ef42d2a59578c5b133d027bed2b9c5bd5ab2690ae7290b5953169f1d
Stable URL of this page: https://metraipolis.com/documento/dc_6f4fb22ffcc2

How to cite

EchoKV: Self-Corrective KV Cache Eviction via Iterative Query Feedback — arXiv:2611.05903 — arXiv (en), https://arxiv.org/abs/2611.05903 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6f4fb22ffcc2 (Metraipolis archive dc_6f4fb22ffcc2, sha256 3936c471ef42d2a59578c5b133d027bed2b9c5bd5ab2690ae7290b5953169f1d)

BibTeX

@misc{dc6f4fb22ffcc2,
  title = {EchoKV: Self-Corrective KV Cache Eviction via Iterative Query Feedback — arXiv:2611.05903},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.05903},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6f4fb22ffcc2; sha256 3936c471ef42d2a59578c5b133d027bed2b9c5bd5ab2690ae7290b5953169f1d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.05903. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

KV cache eviction is not merely a token-ranking problem: the workspace retained for generating eviction queries also determines the quality of those queries. Consequently, an early workspace error can distort later query observations and reinforce an incorrect final eviction decision, particularly when evidence is dispersed across a long context. We present EchoKV, which discovers and addresses this failure mode with iterative correction. It initializes a feedback-compatible bootstrap workspace using proposals at future decode positions, then uses early decode-generated queries to recover the workspace before the remaining lookahead steps. It uses the same number of workspace forward passes as LAQ while adding a tensor re-scoring operation for workspace repair. On LongBench (16 tasks, Mistral-7B-Instruct-v0.2 and Qwen2.5-7B-Instruct), EchoKV shows notable improvements over LAQ on several tasks requiring dispersed evidence, including multi-hop cross-document QA and long single-document comprehension. Across budgets and backbones, EchoKV selectively improves evidence-intensive tasks while maintaining comparable performance on the remaining LongBench subsets. These results identify query--workspace coupling as a practical failure mode of query-driven KV eviction.

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/2611.05903 — 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 2026-11-05T18:12:49.001Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).