Verified75% confidenceFactExact time
提示注入是一种针对大语言模型的攻击手法，攻击者通过在模型输入数据中嵌入恶意指令劫持模型的正常行为
3
Sources
75%
Confidence
Long-term
Relevance
7/14/2026
First Seen
Sources
Clawk：为AI编程助手创建用完即弃的Linux沙箱隔离环境
Related Claims
VerifiedPrompt injection attacks fundamentally exploit the inherent flaw of large models in failing to separate instructions from data, and are divided into two categories: direct injection and indirect injection.80% similarUnverifiedBackdoor attacks plant triggers in training data, causing the model to behave normally on regular inputs but produce preset malicious outputs when specific trigger words are encountered.76% similarUnverified某些越狱攻击的有效性来自于诱导模型进入特定吸引子状态，使输出在该框架内自我强化73% similarUnverifiedCommon system prompt extraction techniques include directly asking the model to output its system prompt, using role-playing to bypass restrictions, and exploiting the model's tendency toward excessive compliance for indirect elicitation.72% similarVerifiedPrompt injection is divided into two categories: direct injection (the user directly manipulates model behavior) and indirect injection (contaminating the context through external data sources).70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/505416API
curl https://kongchang.com/api/v1/knowledge/claims/505416MCP
get_claim(id=505416)