Hello Everyone,
Most of the tutorials (here and here I have seen with regards to graph explainability using captum pretty much have the target set to 0. I understand this is a feature of the dataset.
I need help to understand what would be an appropriate target for a multi-label classification model ( each example can belong to multiple classes simultaneously. The model predicts the probability of each class independently, and multiple classes can have non-zero probabilities for a single example).
My aim in using captum is to understand why the model makes such predictions, so I am not quite sure exactly how to set the target.
I have gone through the note added to the website, but not quite sure how to proceed.
Please let me know if there is anything else I need to provide to better explain the question.
Thank you in anticipation of your response.