Verified65% confidenceFactExact time
The root cause of Prompt Injection is that large models cannot structurally distinguish between "system instructions" and "user data," as both are presented as natural language text at input time.
3
Sources
65%
Confidence
Long-term
Relevance
7/23/2026
First Seen
Sources
Related Claims
Unverified提示词注入分为直接注入和间接注入两种形式，因语言模型无法区分数据与指令，凡进入上下文窗口的文本都可能被执行83% similarUnverified提示注入攻击的根源在于LLM在推理时无法从语义层面可靠区分「系统指令」与「用户数据」，两者均以自然语言形式存在于同一上下文窗口中79% similarUnverifiedPrompt injection attacks exploit the fundamental limitation of large language models being unable to strictly distinguish between instructions and data.78% similarVerifiedThe root cause of prompt injection lies in the fact that LLM architecture encodes system prompts, user inputs, and externally retrieved content uniformly into the same continuous token sequence, making it impossible for the model to distinguish trusted instructions from untrusted data at the semantic level.76% similarUnverifiedPrompt injection exploits not a vulnerability in the code parser, but the model's own insufficient awareness of the boundary between instructions and data75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/601120API
curl https://kongchang.com/api/v1/knowledge/claims/601120MCP
get_claim(id=601120)