Stability of Model Explanations in Interpretable Prototype-based Classification Learning Konferenzpaper