Unverified50% confidenceFactTime unknown

随着上下文窗口从4K扩展到128K及以上，Context Rot问题因信息密度稀释而加剧而非改善

1

Sources

50%

Confidence

Medium-term (~90 days)

Relevance

7/2/2026

First Seen

Valid until: 9/30/2026

Sources

How Multi-Agent Teams Solve AI Hallucination and Make AI Reliable

bilibilililiMozi

Related Claims

Unverified主流模型的上下文窗口已从最初的4K扩展到128K甚至更大，但更大的上下文意味着更高的推理延迟和成本76% similar Unverified即使上下文窗口从4K扩展到128K甚至更长，模型处理长文档时仍会遗漏中间部分信息，研究者称之为「Lost in the Middle」现象73% similar Unverified现代大模型上下文窗口已从早期4K Token扩展到128K甚至更长，注意力机制计算量随上下文长度呈平方级增长73% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/58363

API

curl https://kongchang.com/api/v1/knowledge/claims/58363

MCP

get_claim(id=58363)