ResearchBenchmarksClassification on Overall Across all datasetsClassification on Overall Across all datasetsFollow78AccuracyCARO49.9257.2164.571.79Feb 28, 2026Evaluation resultsMethodMethodLinksAccuracyCohen's KappaCAROPrompting Strategy=Con...Prompting Strategy=Confusion-Aware Rubric Optimization2026.02780.56GradeOptPrompting Strategy=Gra...Prompting Strategy=GradeOpt aggregate feedback2026.02700.47NaivePrompting Strategy=Nai...Prompting Strategy=Naive prompting2026.02510.22