ResearchBenchmarksMetric Correlation with Human Judgment on CT-RATEMetric Correlation with Human Judgment on CT-RATEFollow0.622Pearson CorrelationCT-FineBench0.090560.228530.36650.50447Apr 27, 2026Evaluation resultsMethodMethodLinksPearson CorrelationKendall CorrelationSpearman CorrelationCT-FineBench2026.040.6220.3780.49RaTEScore2026.040.5210.320.434BLEU-22026.040.4950.3530.479BERTScore2026.040.3490.2620.342ROUGE-L2026.040.170.1150.158RadGraph2026.040.1630.1370.171GREEN2026.040.1110.0630.088