ResearchBenchmarksQuestion Answering Calibration on HotpotQA (test)Follow65.7AccuracyLIMECAL62.99663.69864.465.102Oct 14, 2021Evaluation ResultsMethodMethodLinksAccuracyAUCF1@25F1@50F1@75Delta BowLIMECALApproach=LIMECALApproach=LIMECAL2021.1065.778.282.678.473.8—SHAPCALApproach=SHAPCALApproach=SHAPCAL2021.1065.377.8827873.5—BOWPROPApproach=BOWPROPApproach=BOWPROP2021.1064.776.680.376.972.40KAMATHApproach=KAMATHApproach=KAMATH2021.1064.576.880.877.272.8—MAXPROBApproach=MAXPROBApproach=MAXPROB2021.1063.175.779.775.972.2—