Unverified50% confidenceFactExact time
DiacTag提供结构性保证：由于模型不生成新字符序列，输出的基础字符必然与输入保持一致，不可能插入或删除字符
1
Sources
50%
Confidence
Long-term
Relevance
8/5/2026
First Seen
Sources
Related Claims
UnverifiedThe 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.60% similarUnverifiedRegular expressions correspond to regular languages in the Chomsky hierarchy and cannot express arbitrarily deep recursive structures, whereas CFGs correspond to context-free languages and can describe bracket matching, nested structures, and other programming language syntax.58% similarUnverifiedIn code scenarios, slicing by function boundaries preserves more complete semantic units than slicing by fixed character counts.58% similarUnverified传统缓存机制只能匹配完全一致的字符串，导致语义等价的请求被当作全新查询发送给大模型58% similarUnverifiedPrompt engineering has evolved from simple instruction writing into a systematic methodology, and structured prompts (such as those using XML tags or Markdown hierarchies) yield more consistently stable output quality than free-text formats.57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/692413API
curl https://kongchang.com/api/v1/knowledge/claims/692413MCP
get_claim(id=692413)