Verified70% confidenceFactExact time
As the number of conversation turns increases, the weight of early key constraint information in the context window is diluted (attention drift), causing the quality of AI output to gradually decline.
4
Sources
70%
Confidence
Long-term
Relevance
5/28/2026
First Seen
Sources
Harness Engineering: A Practical Guide to AI Programming — From Prompts to Mastery
bilibili图灵诸葛讲Java5/28/2026
Related Entities
Related Claims
VerifiedCurrent large models do not have a true persistent memory mechanism, and as the number of conversation turns increases, the weight of early requirement descriptions in the attention mechanism gradually decays.77% similarUnverified在多轮对话中，冗余输出会占据上下文窗口空间，导致关键信息被挤出窗口（上下文截断问题）77% similarUnverifiedLLM存在上下文窗口限制，在超长对话中面临注意力稀释问题，早期决策记录会被赋予更低注意力权重77% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/2027API
curl https://kongchang.com/api/v1/knowledge/claims/2027MCP
get_claim(id=2027)