Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Xpliquedeel-ai.github.io · Low evidenceXplique is a Python toolkit for explainable AI that aggregates state-of-the-art methods to help users understand complex neural networks. It offers attribution methods (e.g., Saliency, Grad-CAM, Integrated Gradients), feature visualization, concept-based explanations, and metrics, originally built for TensorFlow and with partial PyTorch support. The library also includes an Example-based module that retrieves relevant dataset examples (similar, counterfactuals, prototypes) and provides tutorials, a quick-start guide, and pip installation.
Parse Score