PulseAugur
中
实时 15:45:09
English(EN) When Retrieval Partially Conflicts with Skills: Understanding Procedural Hallucination in Language Models

新的基准 SourceClash 揭示语言模型的“程序性幻觉”

研究人员发现了一种称为“程序性幻觉”的语言模型现象,即模型生成的答案虽然有检索到的上下文支持,但却与其自身的技能或程序相冲突。为了研究这一点,他们创建了一个名为 SourceClash 的基准,包含 1,271 个捆绑包和 10 个程序域中的 6,355 个提示。在九个测试模型中,当检索上下文与技能冲突时,有效答案的比率显著下降至平均 25.7%,而当上下文支持时,该比率则超过 70%。分析表明,针对模型内特定价值路径的干预措施可以提高其最终答案的准确性。 AI

影响 识别出 LLM 中与冲突信息相关的特定故障模式,可能指导未来的模型开发和评估。

排序理由 研究论文,详细介绍了新的基准和对语言模型行为的分析。[lever_c_demoted from research: ic=1 ai=1.0]

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

新的基准 SourceClash 揭示语言模型的“程序性幻觉”

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
研究论文,详细介绍了新的基准和对语言模型行为的分析。[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
6 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]

  1. Hugging Face Daily Papers TIER_1 English(EN) ·

    当检索部分冲突于技能时:理解语言模型的程序性幻觉

    Language models increasingly use skills that provide reusable instructions and procedures for task execution. After a relevant skill has been selected, the model may need to apply its governing procedure together with retrieved task context. When the retrieved context is useful b…