English(EN) What Carries Forward? Pretraining Disclosure Across Language Model Releases

新研究发现，后期发布的语言模型往往会丢失预训练披露信息

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-28 04:00

一项对17个家族的41个语言模型发布进行分析的新研究揭示了一个令人担忧的趋势：后期版本并不总是保留或改进其前代模型的预训练披露。研究人员使用2025年基础模型透明度指数（FMTI）审计数据、处理和计算信息，发现虽然一些模型显示出整体分数有所提高，但这可能掩盖了特定披露指标的丢失。研究结果强调了需要更明确的文档来跟踪连续模型发布中的变化和继承信息的适用性。 AI

影响凸显了后续语言模型发布在透明度方面可能出现的下降，影响了可复现性和信任度。

排序理由分析语言模型预训练披露趋势的学术论文。[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: 17 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 分析语言模型预训练披露趋势的学术论文。[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, policy; 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]

arXiv cs.CL TIER_1 English(EN) · Md Sakib Ul Islam, S M Mehedi Zaman · 2026-10-28 04:00

何以为继？语言模型发布中的预训练披露

arXiv:2610.37465v1 Announce Type: new Abstract: Understanding how language models were pretrained requires information about their data, processing procedures, and computational resources. However, it remains unclear whether successive releases preserve or expand earlier disclosu…

本文如何被排名

报道来源 [1]

何以为继？语言模型发布中的预训练披露

相关实体

相关话题