In this project, we adopt a two-stage approach to textual entailment by first generating human-readable explanations using a T5-based explainer model, and subsequently using those explanations as ...