The BAbI benchmark presents a difficult set of tasks designed to evaluate the skills of AI systems in interpreting commonsense knowledge. It comprises a wide range of situations that require reasoning about everyday notions. By measuring how well AI models can solve these problems, researchers striv