English(EN) Delta N-gram: Forward-Only Continual Writing into a Language Model's Native N-gram Memory

Delta N-gram 实现了对语言模型的连续写入，而无需更改基础参数

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

研究人员开发了一种名为 Delta N-gram 的新颖方法，用于在不更改原始参数的情况下持续更新语言模型。该技术涉及将新信息写入稀疏补丁，保持预训练的 n-gram 表不变。使用此方法的原型证明了即使在多次更新后仍能保留事实的能力，在保留信息和减少模型漂移方面优于传统的基于梯度的更新。 AI

影响该方法可以实现更高效、更稳定的语言模型持续学习，提高其随着时间推移整合新信息的能力。

排序理由详细介绍语言模型更新新方法的学术论文。[lever_c_demoted from research: ic=1 ai=1.0]

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: 4 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 详细介绍语言模型更新新方法的学术论文。[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。

arXiv cs.CL TIER_1 English(EN) · Hao Wu · 2026-10-23 04:00

Delta N-gram: Forward-Only Continual Writing into a Language Model's Native N-gram Memory

arXiv:2610.31125v1 Announce Type: new Abstract: Our previous analysis found that a model's native hashed N-gram table is where continual updates disturb the base least, but its writes needed backpropagation and changed the pretrained table. We present Delta N-gram, a forward-only…