Verified65% confidenceOpinionTime unknown
The root cause of hallucinations in large language models lies in the Transformer architecture's objective of maximizing probability during text generation, rather than prioritizing factual accuracy.
3
Sources
65%
Confidence
Long-term
Relevance
5/31/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedTransformer-based large language models use attention mechanisms to uniformly encode system prompts, user inputs, and externally retrieved content as homogenized token sequences, making it impossible for the model to distinguish trusted instructions from untrusted data at the semantic level.77% similarUnverifiedLarge language models generate fabricated content (hallucinations) with high confidence, including inventing non-existent tool names and forging API return results75% similarUnverified大语言模型在生成查找目标文本时依赖记忆重建而非逐字符复制，在长上下文场景下会出现记忆漂移导致字符串替换静默失败，此现象被称为幻觉性精确75% similarVerifiedThe hallucination problem in large language models is typically divided into two categories: intrinsic hallucination and extrinsic hallucination, where intrinsic hallucination refers to generated content that directly contradicts the input data, and extrinsic hallucination refers to generated information not mentioned in the input and impossible to verify.75% similarVerifiedThe reason large language models fabricate references is that they generate text based on probability rather than database retrieval, a phenomenon known as AI hallucination.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/4600API
curl https://kongchang.com/api/v1/knowledge/claims/4600MCP
get_claim(id=4600)