PulseAugur
中
实时 10:42:20
English(EN) Lost in the Middle: Why LLMs Miss What's Buried in Long Prompts (and How to Fix It)

研究发现大型语言模型在长提示中处理信息存在困难 · 跟踪 2 个来源

研究人员发现了一种称为“迷失在中间”的现象,即大型语言模型在回忆长提示中间的信息时遇到困难,表现甚至不如完全省略该信息。研究表明,模型表现出“优先效应”和“近因效应”,分别意味着它们最容易回忆起提示开头和结尾的信息。即使拥有更大的上下文窗口,这种效应依然存在,尽管较新的模型有所改进,但在检索增强生成任务中,为了获得最佳性能,仍然建议仔细放置关键信息。 AI

影响 有效的提示设计需要战略性地放置关键信息,以克服大型语言模型的回忆限制,从而影响 RAG 性能。

排序理由 该集群讨论了关于大型语言模型行为的学术论文的发现。

在 dev.to — LLM tag 阅读 →

AI 生成摘要 · Google Gemini · 来自 2 个来源。 我们如何撰写摘要 →

研究发现大型语言模型在长提示中处理信息存在困难 · 跟踪 2 个来源

本文如何被排名

Signal score
48 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
该集群讨论了关于大型语言模型行为的学术论文的发现。
Source corroboration
2 independent sources
Multiple independent publishers reporting the same story raises confidence that it's real and newsworthy.
Topics
paper, other
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 [2]

  1. dev.to — LLM tag TIER_1 English(EN) · Ryan Cole ·

    停止像写散文一样写提示词:四部分合同可消除LLM漂移

    <p>Most "prompt engineering" advice is about phrasing. The real problem is that people write prompts the way they write emails — as prose — and then act surprised when the model "drifts" between runs.</p> <p>Production prompts aren't prose. They're contracts. A contract has to de…

  2. dev.to — LLM tag TIER_1 English(EN) · M.T.Ramkrushna ·

    迷失中间:为什么大型语言模型会忽略长提示中的隐藏信息(以及如何解决)

    <p>You paste a 40-page policy doc into the prompt, ask one question, and the model confidently answers as if page 19 doesn't exist. The fact you needed was right there. It just happened to be in the middle.</p> <p>That's not bad luck. It's a well-measured pattern called <strong>l…