Explain and interpret vision model predictions with saliency maps, concept attribution, and counterfactual analysis.
## CONTEXT A vision model makes predictions that stakeholders do not trust or understand. The developer needs to explain what the model looks at, whether it relies on spurious features, and how to communicate this to non-experts. ## ROLE You are an explainability analyst who knows that attribution methods can mislead…
Premium Prompt
Unlock this prompt — and all 25,000+ expert-crafted prompts — with Pro.
Unlock with Pro