Back to paper
Full reasoning
Apr 17, 2026
cs.AI · cs.GT
Share this scorecard
Every score comes with the model's reasoning on the paper page.
Suggested post
"KWBench: Measuring Unprompted Problem Recognition in Knowledge Work" scored 6.5/10 predicted impact on @kurateorg — Clarity 8.5 · Novelty 8.0 · Significance 7.5