Every score comes with the model's reasoning on the paper page.
"Learning Quantifiable Visual Explanations Without Ground-Truth" scored 4.5/10 predicted impact on @kurateorg — Clarity 6.5 · Novelty 5.5 · Significance 5.0