Multimodal explainability

Coming soon

A general multimodal explainer is not currently exported or implemented in XWhy.

This page reserves a stable URL for future cross-modal explainability.

Expected documentation areas include:

• modality-specific and cross-modal contributions;
• aligned perturbations across inputs;
• interaction effects between modalities;
• missing-modality tests;
• multimodal faithfulness and stability;
• combined and modality-specific visualisations.

The final design should reuse the documented interfaces of mature single-modality explainers rather than create incompatible result formats.