Feed Entities Podcasts Deep Dives Developers About

[KongchangAI]

Real-time AI knowledge base. Every claim is traceable to its source, with data freshness under 24 hours.

Powered by 240+ first-hand AI sources

Knowledge

Feed
Hot Entities
Podcasts

Developers

CLI
MCP Server
REST API
About

© 2026 KongchangAI. The freshest, most accurate AI knowledge

Unverified60% confidenceFactTime unknown

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

2

Sources

60%

Confidence

Medium-term (~90 days)

Relevance

7/2/2026

First Seen

Valid until: 9/30/2026

Sources

AI幻觉：大模型为何必然会「胡说八道」及应对策略

bilibili老傅1024

Related Claims

Unverified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73% similar Unverified随着模型上下文窗口从4K到128K再到未来可能的数百万token持续扩大，记忆管理和状态管理的实际效果差距会逐渐缩小72% similar Unverified当前主流模型的上下文窗口已扩展至100K甚至200K tokens，但仍面临Lost in the Middle问题72% similar 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71% 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69% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/55874

API

curl https://kongchang.com/api/v1/knowledge/claims/55874

MCP

get_claim(id=55874)