PulseAugur
中
实时 06:58:10
English(EN) Looping Deeper Without Growing the KV Cache: Recurrent Depth for Causal Encoder-Decoders

新的循环CED架构为大型语言模型大幅削减KV缓存需求

研究人员开发了一种新颖的循环因果编码器-解码器(CED)架构,该架构显著减少了与循环语言模型相关的内存占用。这种新设计在不增加跨层和跨轮次的KV缓存大小的情况下实现了循环深度,而KV缓存大小是当前系统中的一个主要瓶颈。通过策略性地将循环置于编码器中,模型可以迭代地优化全局KV表示,从而提高语言建模性能和长上下文扩展能力。在552B参数的DeepSeek-V4.1-Flash模型上的测试表明,与非循环模型相比,KV缓存使用量大幅减少,在推理过程中实现了显著更高的吞吐量和批处理大小。 AI

影响 这项架构创新可以显著降低推理成本,提高大型语言模型的部署效率,从而支持更广泛的应用和更复杂的场景。

排序理由 学术论文,详细介绍了新的模型架构及其性能改进。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.LG 阅读 →

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

新的循环CED架构为大型语言模型大幅削减KV缓存需求

本文如何被排名

Signal score
37 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
学术论文,详细介绍了新的模型架构及其性能改进。[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
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. arXiv cs.LG TIER_1 English(EN) · Jiaqi Leng, Minshen Zhang, Zhaofeng Sun, Yu-Yang Qian, Peng Zhao, Hao Zhang ·

    无需增大KV缓存即可实现更深层循环:因果编码器-解码器的循环深度

    arXiv:2611.13378v1 Announce Type: new Abstract: Looped language models deepen computation without adding parameters, yet caching full-context key-value (KV) states across layers and rounds severely bounds serving batch size. We design a looped causal encoder-decoder (CED) that ac…