Unverified50% confidenceFactExact time
LLM不理解代码语义，不执行代码、不验证API是否存在、不检查类型是否匹配，只是基于概率分布生成看起来最合理的文本序列
1
Sources
50%
Confidence
Long-term
Relevance
8/5/2026
First Seen
Sources
Related Claims
UnverifiedLLM 在代码任务上的表现往往优于其在纯自然语言推理上的表现，因为代码具有严格语法规则和可验证性74% similarUnverifiedThe 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.68% similarUnverifiedLLM 驱动的代码审查能结合代码语义和上下文进行比规则匹配更智能的判断67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/689013API
curl https://kongchang.com/api/v1/knowledge/claims/689013MCP
get_claim(id=689013)