English(EN) NGM: A Plug-and-Play Training-Free Memory Module for LLMs

新的NGM模块通过无训练记忆增强大语言模型

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-27 04:00

研究人员开发了N-gram Memory (NGM)，一个旨在增强大语言模型 (LLMs) 而无需额外训练的新型模块。NGM利用模型现有的预训练token嵌入作为记忆存储，使用因果N-gram编码器和余弦门控记忆注入器。这种即插即用的方法避免了对独立记忆库或复杂检索系统的需求。NGM已在包括Qwen3和Llama-3.2-Instruct在内的各种基准测试和模型骨干上展示了持续的性能提升，并且对于Qwen3-VL-2B等视觉语言模型也显示出有效性。 AI

影响这个无训练记忆模块可以简化大语言模型的开发，并提高在复杂任务上的性能。

排序理由该集群包含一篇详细介绍增强大语言模型新方法的学术论文。[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: 8 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该集群包含一篇详细介绍增强大语言模型新方法的学术论文。[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, model release; 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: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.AI TIER_1 English(EN) · Yuwen Qu, Wenhui Dong, Camil Hamami, Ming Zhao, Chenyang Si, Caifeng Shan · 2026-10-27 04:00

NGM：LLM 的即插即用免训练内存模块

arXiv:2605.16893v2 Announce Type: replace Abstract: Recent studies have introduced memory-augmented architectures that decouple knowledge storage from neural computation, enabling more direct access to stored representations. However, existing approaches often rely on learned mem…

本文如何被排名

报道来源 [1]

NGM：LLM 的即插即用免训练内存模块

相关实体

相关话题