PulseAugur
中
实时 10:27:32
English(EN) Localization Does Not Imply Steerability: A Mechanistic Study of Sign Errors in LLM Mathematics

大型语言模型数学错误与后期网络符号翻转有关

一项新研究调查了大型语言模型为何在多步数学运算中出现符号错误,发现这些错误是在网络后期发生的,通常在MLP层的最后四分之一内。该研究分析了包括Qwen和Llama 3.3 70B Instruct在内的六个开源模型,并发现了不同的符号错误发生机制,例如固定的“减号戳”层或符号放大器。虽然这些电路的定位在模型之间是一致的,但它们的可控性差异很大,有些模型非常容易被纠正,而另一些模型则表现出高度的冗余,这表明本地化并不保证控制。 AI

影响 识别出大型语言模型在数学推理方面的特定故障模式,可能指导未来模型的开发以提高准确性。

排序理由 学术论文,详细介绍了对大型语言模型行为的机制研究。[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 →

大型语言模型数学错误与后期网络符号翻转有关

本文如何被排名

Signal score
13 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
学术论文,详细介绍了对大型语言模型行为的机制研究。[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
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]

  1. arXiv cs.AI TIER_1 English(EN) · Shradha Agarwal, Tariq J., Deepak Rajbhar ·

    本地化不等于可控性:LLM数学中符号错误机制研究

    arXiv:2611.07478v1 Announce Type: cross Abstract: When large language models do multi-step mathematics, a characteristic failure recurs: the model writes one wrong sign token mid-computation -- a cofactor sign, the minus of a recursive integration-by-parts step -- then executes t…