English(EN) Self-Training Doesn't Flatten Language -- It Restructures It: Surface Markers Amplify While Deep Syntax Dies

研究发现自训练重构语言模型

作者 PulseAugur 编辑部 · [3 个来源] · 2026-05-20 01:44

一篇新的研究论文挑战了对语言模型自训练的普遍理解，认为它重构而非扁平化语言。研究发现，虽然话语标记等表面语言特征有所增加，但问句和被动语态等深层句法结构却有所下降。这种“结构深度假说”认为，语言特征的衰减率主要由其结构复杂性决定，而不仅仅是模型输出中的频率。 AI

影响揭示了自训练以复杂的方式改变语言模型输出，影响数据策展和LLM文本检测。

排序理由该集群包含一篇详细介绍语言模型行为新发现的研究论文。

AI 生成摘要 · Google Gemini · 来自 3 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: Research; 该集群包含一篇详细介绍语言模型行为新发现的研究论文。
Source corroboration: 3 independent sources; Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
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: 104 days old; Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.
Coverage growth since scoring: +1 source(s) since last score; New sources have picked up this story since our last re-score. Score will update on the next scoring pass.

完整方法见我们的编辑标准。

arXiv cs.AI TIER_1 English(EN) · Ming Liu · 2026-05-22 04:00

自训练不使语言扁平化——它重构语言：表面标记被放大，深层句法消亡

arXiv:2605.20602v1 Announce Type: cross Abstract: Successive self-training on a language model's own outputs is widely characterized as a process of flattening: diversity drops, distributions narrow, and the text becomes "more like itself." We provide evidence that this character…
arXiv cs.CL TIER_1 English(EN) · Ming Liu · 2026-05-20 01:44

自训练不使语言扁平化——它重构语言：表面标记被放大，深层句法消亡

Successive self-training on a language model's own outputs is widely characterized as a process of flattening: diversity drops, distributions narrow, and the text becomes "more like itself." We provide evidence that this characterization is incomplete. Across eleven generations o…
Hugging Face Daily Papers TIER_1 English(EN) · 2026-05-20 01:44

自训练不使语言扁平化——它重构语言：表面标记被放大，深层句法消亡

Successive self-training on a language model's own outputs is widely characterized as a process of flattening: diversity drops, distributions narrow, and the text becomes "more like itself." We provide evidence that this characterization is incomplete. Across eleven generations o…