English(EN) Is Convergence Inevitable? Tracing Output Homogeneity Back to Base Models

LLM输出同质性可能源于预训练，而非仅仅是调优

作者 PulseAugur 编辑部 · [1 个来源] · 2026-08-13 04:00

一项新的研究论文表明，在大型语言模型（LLM）输出中观察到的同质性可能源于预训练阶段，而非仅仅在调优过程中产生。研究发现，语义收敛可以通过监督微调（SFT）被揭示和放大，但不能由其引入。此外，通过单独提示就可以在基础模型中诱导类似指令的崩溃，这表明收敛可能是LLM训练目标的自然结果，使得后调优干预不足以缓解。 AI

影响表明当前的调优技术可能不足以解决LLM输出同质性问题，需要重新评估预训练策略。

排序理由在arXiv上发表的研究论文，讨论LLM训练动态。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 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; 在arXiv上发表的研究论文，讨论LLM训练动态。[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: 53 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.CL TIER_1 English(EN) · Alexandrine Fortier, Hazel Chen, Peter West · 2026-08-13 04:00

融合是必然的吗？追溯输出同质性到基础模型

arXiv:2608.11426v1 Announce Type: new Abstract: The lack of diversity in LM content is widely attributed to the alignment process, but how and where exactly in the pipeline this collapse begins is unknown. We argue that output homogeneity is likely learned during the pretraining …

本文如何被排名

报道来源 [1]

融合是必然的吗？追溯输出同质性到基础模型

相关实体

相关话题