ResearchBenchmarksTrait-level Essay Scoring on LAILA (test)Follow0.411RelevanceMOOSE0.285160.317830.35050.38317Apr 19, 2026Evaluation ResultsMethodMethodLinksRelevanceQWKOrganizationVocabularyStyleDevelopmentMechanicsGrammarHolistic ScoreAverage ScoreMOOSEModel=MOOSEModel=MOOSE2026.040.411—0.6270.6420.6490.5850.5860.6230.6490.597XGBModel=XGBModel=XGB2026.040.36—0.6450.6410.6410.5830.5770.6190.6790.593MAPLEModel=MAPLEModel=MAPLE2026.040.29—0.6470.6860.7020.6070.6340.6640.7230.619