Do Explanations make VQA Models more Predictable to a Human?

Open in new window