Detecting Stealth Sycophancy in Mental-Health Dialogue with Dynamic Emotional Signature Graphs

ArXi:2605.03472v1 Announce Type: cross As conversational AI therapists are increasingly used in psychological settings, reliable offline evaluation of therapeutic response quality remains an open problem. This paper studies multi-domain -dialogue evaluation without relying on large language models as final judges. We use a direct LLM judge as a baseline that reads raw dialogue text and predicts whether the target response is harmful, productive, or neutral.