PulseAugur
中
实时 09:20:17
English(EN) From Unreliable to Usable: Decomposing Failure Modes and Confidence Signals in LLM Entity Extraction - Evidence from Cyber Threat Intelligence

研究发现大型语言模型实体提取不可靠

一篇新的arXiv论文研究了大型语言模型(LLMs)在信息提取方面的可靠性,特别是在网络威胁情报文档方面。研究发现,即使是像GPT-6.1-Sol、DeepSeek-V4 Flash和GLM-5.3-Flash这样先进的模型也表现出显著的不可靠性,其精确率低且校准误差高。与普遍假设相反,主要的故障模式不是幻觉,而是背景提及和范围不正确,而实际不支持的候选词很少见。虽然置信度信号有助于优先审查,但它们不能保证准确性,因为即使是微小的句子扰动也会极大地改变高置信度的判断。 AI

影响 强调了大型语言模型数据提取的局限性,建议在高风险应用(如威胁情报)中要谨慎。

排序理由 详细介绍大型语言模型评估和故障模式的学术论文。[lever_c_research降级: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 →

研究发现大型语言模型实体提取不可靠

本文如何被排名

Signal score
14 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
详细介绍大型语言模型评估和故障模式的学术论文。[lever_c_research降级: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.CL TIER_1 English(EN) · Jiahong Qian ·

    从不可靠到可用:LLM实体提取中的失败模式和置信度信号分解——来自网络威胁情报的证据

    arXiv:2610.20417v1 Announce Type: new Abstract: Audits of large language model (LLM) extraction from full-length documents report severe unreliability, leaving two questions open: what the failures actually consist of, and whether confidence signals can turn unreliable output int…