Unverified50% confidenceFactExact time
Large language models have been shown to be capable of influencing human beliefs
1
Sources
50%
Confidence
Long-term
Relevance
10/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe paper "Framing the Narrative: Ideological Mimicry in Large Language Models" argues that the political stance of large language models is not fixed, but shifts systematically in response to the way users express themselves.73% similarUnverified论文提出为防止大语言模型声称拥有意识而进行的安全微调会意外改变模型对'心智'的整体表征，扭曲其对人类信仰与价值观的建模73% similarUnverifiedIntroducing large language models to give the system semantic understanding capabilities enables it to understand the deceptive differences in phrasing between scam scripts and system warnings.72% similarUnverifiedLarge language models have a sycophancy characteristic, tending to organize their output according to the inclination of the prompt rather than reflecting factual truth.71% similarUnverified2023年多项研究指出，大语言模型的使用可能影响用户的元认知能力，即人对自身认知过程的监控和调节能力71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/963950API
curl https://kongchang.com/api/v1/knowledge/claims/963950MCP
get_claim(id=963950)