Unverified50% confidenceFactExact time
LLMs do not maintain a factual database and lack the metacognitive ability to know 'whether an operation was actually performed'; when the context strongly implies a certain output pattern, they generate text conforming to that pattern with high confidence, even if the content is entirely fabricated.
1
Sources
50%
Confidence
Long-term
Relevance
8/9/2026
First Seen
Sources
AI幻觉实录：Perplexity假装登录了你的账户
Related Claims
UnverifiedLLM architectures cannot strictly distinguish between 'instructions' and 'data' at the semantic level, which is the root cause of why prompt injection is difficult to fundamentally eliminate.77% similarVerifiedHallucination problems in LLM code generation manifest as fabricating non-existent APIs, incorrectly combining parameter orders of real APIs, or generating deprecated usage based on outdated documentation.71% similarUnverifiedLLM幻觉问题源于模型基于概率预测生成文本而非检索真实数据，在涉及具体统计数字、引文来源和法律条款时尤为危险66% similarUnverified直接让 LLM 生成 SQL 风险很高，因为模型不了解表结构、字段含义和业务定义，容易生成语法正确但结果错误的查询66% similarUnverifiedLLM在训练过程中将文本关联保存为数学权重参数，并彻底删除源数据，因此不会搜索内部数据库来复制答案65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/717057API
curl https://kongchang.com/api/v1/knowledge/claims/717057MCP
get_claim(id=717057)