Beyond Accuracy: Counterfactual Fragility and Demographic Bias in Clinical Evaluation of LLMs
Findings show that accuracy, CFV, answer stability, EDD, and stereotype evidence capture different evaluation aspects, and that correct answers do not guarantee reliable performance on the counterfactual validity task.
Chaitai Deb Purkayastha, B. Bolla, Vishnu Surya Reddy Nandi
· 0 citations