Unverified75% confidenceOpinionTime unknown
大语言模型（LLM）的普及压缩了信息生产、筛选和传播的专业门槛，削弱了传统守门人（媒体、学术机构、专家）的权威
1
Sources
75%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe context windows of current mainstream large language models have expanded to hundreds of thousands or even millions of tokens, but under ultra-long contexts the model's attention weights on earlier information decay significantly (the "lost in the middle" problem).68% similarVerified大语言模型在长时间对话中会因上下文窗口限制而丢失早期信息67% similarUnverifiedThe current hotspot in interpretability research has clearly shifted toward large language models (LLMs), with relatively limited attention given to visual convolutional networks.67% similarUnverified在多轮对话场景中，随着历史消息累积，系统指令相对业务内容的比重下降，格式遵从度随之降低66% similarUnverified大语言模型存在三大缺陷：知识陈旧（存在截止日期，无法获取实时数据）、缺乏私有数据、过于被动65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/2561API
curl https://kongchang.com/api/v1/knowledge/claims/2561MCP
get_claim(id=2561)