Logical Reasoning with Span-Level Predictions for Interpretable and Robust NLI Models

Joe Stacey, Pasquale Minervini, Haim Dubossarsky, Marek Rei