← 研踪 PaperTrailACL Anthology · ACL 2026:acl:2026.acl-long.859 · 记录 #1
SpecCache: Speculative KV Cache Reuse for Efficient RAG Serving
来源日期（仅到月份） 2026-07
来源记录日期 2026-07
雷达观测 2026-11-01 21:58
摘要
Retrieval-Augmented Generation (RAG) significantly enhances LLMs but faces high prefill latency during long-context processing. While KV cache reuse can mitigate this, current methods relying on shallow features or static heuristics often fail to identify critical tokens for recomputation, resulting in generation quality degradation.We have an insight that KV deviations are more pronounced in deep layers.However, directly extracting deep-layer features from the target model is computationally prohibitive. Crucially, we find that the deep-layer features of a lightweight speculative model exhibit strong consistency with the target model in the selection of critical tokens for recomputation.In light of these insights, we propose SpecCache, which employs deep-layer hidden-state norms from a speculative model as a proxy to guide the critical token selection for target large model.Experiments demonstrate that SpecCache outperforms state-of-the-art (SOTA) baselines. Compared to full KV recomputation, it reduces time-to-first-token (TTFT) by 2.17-3.95\times and increases inference throughput by 2.7-5.2\times, with negligible degradation in generation quality relative to full recomputation.
论文发现出处
这些页面提供论文线索，不代表同行评审或质量认证。
观测到的版本
平台记录号代表本站观测到的元数据变化，不是作者修订号。来源日期仅到月份或年份时，按该月或该年首日存档。
记录 #1 · 2026-07 · 最新观测版本
SpecCache: Speculative KV Cache Reuse for Efficient RAG Serving
Retrieval-Augmented Generation (RAG) significantly enhances LLMs but faces high prefill latency during long-context processing. While KV cache reuse can mitigate this, current methods relying on shallow features or static heuristics often fail to identify critical tokens for recomputation, resulting in generation quality degradation.We have an insight that KV deviations are more pronounced in deep layers.However, directly extracting deep-layer features from the target model is computationally prohibitive. Crucially, we find that the deep-layer features of a lightweight speculative model exhibit strong consistency with the target model in the selection of critical tokens for recomputation.In light of these insights, we propose SpecCache, which employs deep-layer hidden-state norms from a speculative model as a proxy to guide the critical token selection for target large model.Experiments demonstrate that SpecCache outperforms state-of-the-art (SOTA) baselines. Compared to full KV recomputation, it reduces time-to-first-token (TTFT) by 2.17-3.95\times and increases inference throughput by 2.7-5.2\times, with negligible degradation in generation quality relative to full recomputation.
查看此版本解读 →查看此版本原文 ↗