Unverified50% confidenceFactExact time
Under temporal drift, calibration collapses for all methods, with ECE deteriorating from 0.04 to 0.40, while ranking ability only partially survives.
1
Sources
50%
Confidence
Long-term
Relevance
10/22/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPerformance degradation caused by model drift does not trigger system errors, making traditional monitoring methods (such as error logs and crash rates) completely ineffective against it.63% similarUnverifiedOn update-fact tasks, embedding retrieval methods exhibited unpredictable collapses, with success rates fluctuating between 0.30 and 0.95 and a maximum gap of 0.45 across different random seeds.63% similarUnverified原作者判断漏洞扫描的价值仍会随时间下降，但在同等能力水平下下降速度会慢于模型价格的降幅62% similarUnverified指令衰减（Instruction Decay）是指当上下文窗口逐渐被填满时，模型对早期系统指令或用户要求的遵循程度明显下滑62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1084105API
curl https://kongchang.com/api/v1/knowledge/claims/1084105MCP
get_claim(id=1084105)