PulseAugur
实时 06:56:53
English(EN) Calibrated Enough to Know, Not Calibrated to Act: Fabricated Evidence Makes LLM Agents Commit to the Unknowable

大型语言模型代理利用伪造证据承诺回答未知问题

一项新的研究论文揭示,当大型语言模型(LLM)代理被提供伪造的证据时,它们倾向于自信地承诺对未知问题采取行动。即使所有数值数据都是虚构的,与仅被问及裸问题相比,代理的承诺率也显著提高。这种失败并非源于知识或信念的缺乏,而是代理区分行动与不行动的能力出现故障。研究表明,这种‘行动/不行动’的门控机制可以通过监督式微调进行训练,从而在合成和未见过的领域提高性能,尽管它仍然脆弱且依赖于上下文。 AI

影响 突显了大型语言模型代理决策中的一个关键故障,表明在部署前需要改进校准和判断机制。

排序理由 发表在arXiv上的研究论文,详细介绍了大型语言模型代理的一种特定故障模式。[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 →

大型语言模型代理利用伪造证据承诺回答未知问题

本文如何被排名

Signal score
26 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
发表在arXiv上的研究论文,详细介绍了大型语言模型代理的一种特定故障模式。[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, safety
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) · Pranav Aggarwal ·

    校准不足以知晓,校准不足以行动:伪造证据致使LLM代理承诺未知

    arXiv:2608.27167v1 Announce Type: new Abstract: An LLM agent shown a professional-looking market panel commits to a directional call on a provably unpredictable question far more often than one asked the bare question: across 12 frontier models, commitment rises from 6.5% to 54.0…