Metraipolis › archive › dc_3fcf75219013
BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference — arXiv:2609.04971
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.04971
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T22:36:01.042Z
- SHA-256 of the archived text
4f5b8463a99c7111ff477e9355a857d906feed6f0d1b7548bcf26918fb5e9b39- Stable URL of this page
https://metraipolis.com/documento/dc_3fcf75219013
How to cite
BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference — arXiv:2609.04971 — arXiv (en), https://arxiv.org/abs/2609.04971 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_3fcf75219013 (Metraipolis archive dc_3fcf75219013, sha256 4f5b8463a99c7111ff477e9355a857d906feed6f0d1b7548bcf26918fb5e9b39)
BibTeX
@misc{dc3fcf75219013,
title = {BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model Inference — arXiv:2609.04971},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.04971},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_3fcf75219013; sha256 4f5b8463a99c7111ff477e9355a857d906feed6f0d1b7548bcf26918fb5e9b39},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.04971. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Large Reasoning Models (LRMs) achieve superior problem-solving through extended Chain-of-Thought (CoT) generation, but the resulting key-value (KV) cache grows linearly with sequence length and creates severe memory bottlenecks, often exceeding GPU capacity for long reasoning traces. Existing KV cache compression methods rely on recent queries to estimate future token importance, implicitly assuming these serve as reliable proxies for future attention patterns. We demonstrate that this assumption fails in long-horizon reasoning: certain decoding steps generate Thought Revisiting Tokens (TRT) that re-attend to distant previous context, such as task-solving plans formulated early in the trace. Through systematic analysis, we discover that queries corresponding to the TRT cluster into a small number of similarity groups in the embedding space. Based on this insight, we propose BeaconKV, a training-free KV cache compression method that maintains beacon queries, compact representatives for each global query cluster, to anticipate which KV pairs will be revisited without storing the entire query history. Across four open-source LRMs and diverse reasoning benchmarks, BeaconKV generally outperforms existing compression methods, achieving up to $5.8\times$ memory reduction while nearly preserving full cache accuracy and improving throughput by over $4.3\times$. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-07 in cs.CL). Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.04971 — 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 2026-10-16T22:21:16.323Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).