Tamp-X: Attacking explainable natural language classifiers through tampered activations

Publication
Computers & Security