Robustness Evaluation is a current AI evaluation practice associated with AI safety field in Artificial Intelligence. It supports Artificial Intelligence development, governance, evaluation, operations, or security.
Evaluation of model behavior under variation, stress, or distribution shift.