1 paper
Stephen McAleese, Mark Keane
Counterfactual explanations can be used to interpret and debug text classifiers by producing minimally altered text inputs that change a classifier's output. In this work, we evalu…