English(EN) RACE: RoPE-Aligned Attention Compression Based on Query-Key Activation Energies

新的RACE方法压缩LLM注意力，实现更快、更高效的推理

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-21 04:00

研究人员开发了RACE，一种通过降低查询和键向量维度的创新方法来压缩大型语言模型中注意力机制的方法。该技术基于联合查询-键激活能，允许模型在没有运行时重建或专用内核的情况下保留原生的降维表示。RACE在推理效率方面取得了显著的改进，减少了FLOPs、首个token的生成时间和GPU内存使用量，同时在各种任务上保持了具有竞争力的性能。 AI

影响通过优化注意力机制，降低LLM的推理成本并提高吞吐量。

排序理由详细介绍LLM注意力压缩新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.AI 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 0 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍LLM注意力压缩新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, infra; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: 3 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Irene Tenison, Abdulrahman Alabdulkareem, Daniel Karl I. Weidele, Nima Dehmamy, Dan Gutfreund · 2026-10-21 04:00

RACE：基于查询-键激活能的RoPE对齐注意力压缩

arXiv:2610.28385v1 Announce Type: new Abstract: Head-dimension compression can reduce the memory required by the key-value (KV) cache, but existing methods often reconstruct full-dimensional representations during inference, adding significant decoding FLOPs overhead. We formaliz…

本文如何被排名

报道来源 [1]

RACE：基于查询-键激活能的RoPE对齐注意力压缩

相关实体

相关话题