Python Libraries for Explainable AI logopeterindia.net
Knowledge Hub

Python Libraries for Explainable AI (XAI): SHAP, LIME, Captum & More

Python toolkits for making machine learning models interpretable — attribution, counterfactuals, rules and evaluation.

24 topics•5 categories•Updated November 2026

Feature Attribution

5 entries

Explain individual predictions.

01SHAPGame-theoretic Shapley value explanations for any model. 02LIMELocal interpretable model-agnostic explanations; last commit on the default branch was in July 2021. The old Papers with Code page now redirects elsewhere. 03CaptumModel interpretability library for PyTorch. 04ELI5Visualize and debug machine learning models with a unified API. 05BreakDownDecomposes a prediction into feature contributions.

Toolkits & Dashboards

6 entries

Broader explainability suites.

06InterpretMLMicrosoft toolkit for glass-box models such as the explainable boosting machine, and black-box explainers. 07AI Explainability 360 (AIX360)IBM-originated toolkit, now Linux Foundation, with many explainability methods; last GitHub update September 2026. 08OmniXAISalesforce library with a unified interface for tabular, text, image and time series; last GitHub update June 2026. 09Alibi ExplainSeldon library for model inspection and explanations. 10ShapashMAIF dashboards that make SHAP and LIME results readable. 11XAI (Institute for Ethical AI)Library for bias detection and feature importance.

Rules, Counterfactuals & Text

3 entries

Other explanation styles.

12AnchorsIF-THEN rule explanations; last GitHub update July 2022, so treat as a reference implementation. 13DiCEDiverse counterfactual explanations. 14interpret-textExplainers for text models; last GitHub update February 2024.

Evaluating Explanations

1 entries

Check that explanations are trustworthy.

15QuantusToolkit for evaluating neural network explanations.

XAI Libraries

2 entries

Python explainability libraries.

16SHAPShapley-value explanations for models. 17InterpretMLGlass-box models and explainers.

Related Pages

7 entries

Neighbouring pages on this site.

18PythonPython hub. 19XAI FrameworksExplainable AI frameworks. 20Machine Learning FrameworksML frameworks. 21Deep Learning FrameworksDeep learning frameworks. 22Python Libraries for Data ScientistsCore data science libraries. 23AI Explainability Techniques, Tools and LibrariesTechniques, mechanistic interpretability, evaluation of explanations, guidance and libraries. 24Artificial IntelligenceHub of AI platforms, models, frameworks, governance and learning resources.

Frequently Asked Questions

What is explainable AI?

Explainable AI is the set of methods that make a machine learning model's predictions understandable to people, for example by showing which features drove a result.

Which XAI library should I start with?

SHAP and LIME are the most widely used for tabular, text and image models; Captum is a natural choice for PyTorch models and InterpretML for glass-box models.

Are all these libraries actively maintained?

No. Several have had no recent commits (LIME since 2021, Anchors since 2022, interpret-text since 2024); the notes give last-commit dates from October 2026.

Which additional references does this page add?

The section 'XAI Libraries' lists SHAP, Captum, InterpretML, LIME.

How current is this page?

Links were checked in October 2026. Sites that block automated checks are kept and may need a manual click. Last-commit dates are from GitHub in October 2026 and show activity only; check each project's release notes before adopting it.