Measuring Delusion Reinforcement Across the Directness-Need Spectrum
Separating affective support from epistemic sycophancy in conversational models
Image will load when scrolled into view
LLM SafetyEpistemic SycophancyBehavioral Evaluation
About the Project
Building evaluations of when conversational language models reinforce ungrounded beliefs as user cues become more indirect.
The benchmark separates affective support from epistemic sycophancy and tests conversational responses across autonomy, belonging, and safety needs. This is an ongoing model-evaluation project.
Project Details
StatusOngoing Research
Role
Research Fellow, Sentient Futures Project Incubator
Stack
LLM Evaluation
Behavioral Benchmarks
Controlled Comparisons