Community-Aligned Behavior Under Uncertainty: Evidence of Epistemic Stance Transfer in LLMs

Patrick Gerard, Aiden Chang, Svitlana Volkova

When large language models (LLMs) are aligned to a specific online community, do they exhibit generalizable behavioral patterns that mirror that community's attitudes and responses…