Stability of Personal Value Expression in Large Language Models Across Diverse Contexts

Large Language Models exhibit varying degrees of stability in their expression of personal values across different conversational contexts, with some models and model families demonstrating higher value stability than others.