English(EN) Context Engineering: Drift, Bloat, and Lost Attention

AI上下文工程面临“中间丢失”效应的挑战

作者 PulseAugur 编辑部 · [1 个来源] · 2026-09-11 13:01

上下文工程，区别于提示工程，专注于管理模型在推理时接收的所有输入，包括系统提示、工具定义和消息历史。一个关键挑战是“上下文腐烂”，即随着上下文长度的增加，模型变得不那么可靠，这不是由于内存故障，而是因为Transformer的注意力机制被过度拉伸。研究表明存在“中间丢失”效应，模型难以回忆起长上下文中间部分的信息，而在信息位于开头或结尾时表现最佳。 AI

影响强调了当前LLM上下文窗口利用的局限性，表明需要改进上下文管理技术，而不仅仅是增加窗口大小。

排序理由该条目讨论了关于LLM上下文窗口限制和“中间丢失”效应的研究结果，并引用了学术研究。[lever_c_demoted from research: ic=1 ai=1.0]

在 Towards AI 阅读 →

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

本文如何被排名

Signal score: 31 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 该条目讨论了关于LLM上下文窗口限制和“中间丢失”效应的研究结果，并引用了学术研究。[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, 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 [1]

Towards AI TIER_1 English(EN) · M. Haseeb Hassan · 2026-09-11 13:01

上下文工程：漂移、臃肿和注意力丧失

<p>A coding agent that correctly diagnoses a failing test at turn three will sometimes, by turn forty, suggest reverting the fix it already applied. It has read the same log file five times, inherited every tool result along the way, and none of that volume made it more reliable.…

本文如何被排名

报道来源 [1]

上下文工程：漂移、臃肿和注意力丧失

相关实体

相关话题