Unverified50% confidenceOpinionExact time
Current large language models tend to apply the same set of expression habits to all contacts when generating replies, lacking awareness of relationships.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/17/2026
First Seen
Valid until: 2/15/2027
Sources
Related Entities
Related Claims
Unverified同一个大型语言模型在使用不同语言交流时,其语气、态度和礼貌程度可能存在显著差异77% similarVerifiedLarge language models have a sycophancy characteristic, tending to organize their output according to the inclination of the prompt rather than reflecting factual truth.77% similarUnverifiedThe researchers point out that large language models often struggle to strike a balance between empathetic interaction and adherence to cognitive stimulation therapy norms.77% similarUnverifiedLong's Interaction Hypothesis emphasizes that receiving corrective feedback during meaningful interaction is a key mechanism for language development.76% similarVerifiedLarge language models exhibit a sycophancy tendency, changing originally correct positions with relatively high probability when users express dissatisfaction.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1178617API
curl https://kongchang.com/api/v1/knowledge/claims/1178617MCP
get_claim(id=1178617)