Tag
This paper introduces Counterfactual Explanation Consistency (CEC), a framework to detect and mitigate hidden procedural bias in outcome-fair models by aligning feature attributions between individuals and their counterfactual counterparts, with experiments on credit and income datasets.