No convergence for IH for larger input strings about path_explain HOT 2 OPEN

This is odd! I didn't have this problem running on a transformer with longer sentences and far fewer interpolation points (like 500 or so). Could you share some of the code you used to generate this plot? When you generate interaction values for NLP models, you have to use an explainer tailored towards handling embedding-based inputs and be careful about where sums are taking place.

To be clear: convergence is theoretically guaranteed for IH values in the same way it is guaranteed for IG values.

from path_explain.

For example: could you try comparing your code to https://github.com/suinleelab/path_explain/blob/master/examples/natural_language/transformers/check_completeness_embeddings.ipynb

from path_explain.