Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
XAI-Deepfakes Explainability Toolkitgithub.com · Low evidenceXAI Deepfakes is a software framework that quantitatively evaluates explainable AI methods for deepfake detection using an EfficientNet-based model. It assesses explanation approaches like GradCAM++, RISE, SHAP, LIME, and SOBOL by testing their ability to identify regions in fake images that influence detector decisions through adversarial modifications.
Parse Score