Unverified50% confidenceFactExact time
Large language models typically trigger built-in safety alignment mechanisms when facing topics such as political elections and evaluations of candidates.
1
Sources
50%
Confidence
Long-term
Relevance
10/28/2026
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.72% similarUnverified大语言模型在生成摘要时天然倾向于反映评论中的多数意见，可能将低频的安全警告边缘化71% similarUnverifiedMultiple academic tests have found that mainstream large language models generally exhibit a tendency toward center-left positions in political orientation tests (such as the Political Compass Test).71% similarUnverifiedSafety-aligned language models have a one-dimensional direction in activation space associated with refusing harmful requests.69% similarVerifiedLarge language models have a sycophancy characteristic, tending to organize their output according to the inclination of the prompt rather than reflecting factual truth.68% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1109514API
curl https://kongchang.com/api/v1/knowledge/claims/1109514MCP
get_claim(id=1109514)