dalex

The core `explain()` function creates a consistent interface for any predictive model, enabling standardized analysis across frameworks like scikit-learn, xgboost, and keras, as highlighted in the README's integration examples.

Offers both local (e.g., LIME) and global (e.g., partial dependence) techniques, providing detailed insights into individual predictions and overall model behavior, supported by cheatsheets and tutorials in the resources.

Available as R and Python packages, catering to diverse teams and integrating with popular ML ecosystems, which is emphasized in the installation and overview sections.

Includes Arena, an interactive dashboard for model comparison, enhancing exploratory analysis and stakeholder communication, as noted in the Python package features.

Provides dedicated modules to assess model fairness, aligning with responsible AI practices and addressing ethical concerns, which is a key feature highlighted in the description.