Unverified60% confidenceFactExact time
A 2023 Stanford University paper titled 'Lost in the Middle' found that models pay significantly less attention to information in the middle of the context window than at the beginning and end, exhibiting a U-shaped attention distribution.
2
Sources
60%
Confidence
Long-term
Relevance
7/11/2026
First Seen
Sources
The Three Eras of AI Agents: The Capability Leap from Tool Calling to Autonomous Orchestration
youtubeTheo - t3․gg7/10/2026
Related Claims
Verified研究表明当上下文过长时，模型对中间部分信息的注意力会显著下降，即"Lost in the Middle"现象83% similarUnverifiedAn "lost in the middle" effect exists in ultra-long contexts, meaning the model has stronger memory for content at the beginning and end of a sequence while attention to middle passages decreases; this phenomenon was named by Stanford researchers.79% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/488615API
curl https://kongchang.com/api/v1/knowledge/claims/488615MCP
get_claim(id=488615)