Unverified50% confidenceOpinionExact time
Without enterprise business semantic knowledge, large language models can only rely on field names for inference, which is highly prone to semantic drift.
1
Sources
50%
Confidence
Long-term
Relevance
10/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedVector embedding models rely on general language distributions; internal enterprise abbreviations, project codenames, and personal name shortcuts are nearly absent from pre-training corpora, causing semantic similarity calculations to be inaccurate.79% similarUnverifiedLarge language models can generate syntactically correct query statements, but cannot automatically understand the definitions of specific business terms within an enterprise (such as the precise meanings of "active users," "gross margin," and "North region sales").79% similarVerifiedText generated by large language models tends to have low perplexity, with word choices that are overly consistent with probabilistic expectations, lacking the unexpectedness of human writing.73% similarUnverified大语言模型在涉及企业内部文档、专业领域资料或最新数据时容易产生幻觉（一本正经地胡说八道）73% similarUnverifiedA language model saying a model name does not mean it has actually read the relevant announcement — it may have guessed correctly or misremembered.73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/974859API
curl https://kongchang.com/api/v1/knowledge/claims/974859MCP
get_claim(id=974859)