interpretml/interpret-text
An open-source library for explaining and visualizing blackbox text classification models using state-of-the-art interpretability techniques.

Interpret-Text builds on the Interpret ML framework to extend interpretability capabilities to NLP and text-based machine learning models. It provides a unified SDK and Jupyter notebooks that let users run experiments across multiple explainers (glass-box, grey-box, and black-box approaches) and compare results visually. The library supports both global explanations across labels and local explanations for individual documents.