Neural Network Interpretability

Is there anything in this library that would be useful in interpreting chemical graph neural networks?

@jmarsh, some of this will certainly be useful in pathology WSI.