Unverified50% confidenceFactExact time
当底层大语言模型进行版本更新时，即使提示词不变，输出分布也可能发生变化
1
Sources
50%
Confidence
Long-term
Relevance
8/12/2026
First Seen
Sources
Related Claims
Unverified大语言模型本质上是概率性的，相同的输入可能产生不同的输出，这与传统软件的确定性行为截然不同75% similarUnverifiedLarge language models exhibit an 'instruction drift' phenomenon, where the model's probability of following early instructions gradually decreases as conversation turns increase and the context window is filled with large amounts of code.74% similarUnverifiedLarge language models have inherent non-determinism — even with identical prompt inputs, the model may return results that differ in format, content, or even logic at different times.72% similarUnverified系统提示词的语言选择会影响输出风格，当系统提示词语言与用户输入语言不一致时会产生行为叠加效应69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/735058API
curl https://kongchang.com/api/v1/knowledge/claims/735058MCP
get_claim(id=735058)