Unverified50% confidenceSolutionExact time
Negative prompting is often ineffective for language models, because negated concepts actually activate the model's attention.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Verified提示注入的根本原因在于大语言模型缺乏区分指令层与数据层的原生机制73% similarUnverified元评论是当前大语言模型的一个典型系统性缺陷，指模型在生成内容时混入对自身生成行为的描述性语言72% similarUnverified大语言模型在生成摘要时天然倾向于反映评论中的多数意见，可能将低频的安全警告边缘化72% similarUnverifiedPerplexity measures a language model's ability to predict test text; the lower the value, the less surprised the model is by the text.71% 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70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/909878API
curl https://kongchang.com/api/v1/knowledge/claims/909878MCP
get_claim(id=909878)