Every score comes with the model's reasoning on the paper page.
"ChemVA: Advancing Large Language Models on Chemical Reaction Diagrams Understanding" scored 7.0/10 predicted impact on @kurateorg — Clarity 7.5 · Significance 7.0 · Novelty 6.5