ResearchDatasetsHotpot QAHotpot QAFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsmulti-hop question answeringHOTPOT-QA68.8Accuracy20Open-domain question answeringHotpot QA (val)24.6EM8