Expired60% confidenceFactExact time
Claude's context window can reach 200K tokens, but as the conversation accumulates, the attention weights of early information gradually decay, which is the phenomenon known in academia as 'Lost in the Middle'.
2
Sources
60%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Mastering Claude Code: Five Hardcore Strategies to Transform It from Intern to Architect
bilibiliAISolo大西瓜6/21/2026
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73% similarUnverifiedA 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.67% similarUnverifiedWhen Claude Code's context window approaches capacity, earlier conversations get compressed or discarded67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/41839API
curl https://kongchang.com/api/v1/knowledge/claims/41839MCP
get_claim(id=41839)