I have created a model that classifies sentences into different categories using embedding layer. I would like to calculate the salient token (the input token that was the most
