I find your method for interpretability of sequence models very interesting. I would like to apply this approach to a medical dataset aiming at having both the accuracy and interpretability. I was wondering are there any updates of the code availability.