Announcement
New paper examines whether LLM political stances shift with framing
A researcher sharing the paper says the same model can show different stances under different framing.
TLDR
A researcher announced “Framing the Narrative: Ideological Mimicry in Large Language Models,” a paper studying whether LLM political views are fixed or change systematically with an interaction. The researcher sums it up as: “Same model. Different framing. Different stance.”
Combined views
302
2 Sources, first seen ago