ResearchBenchmarksMultiple-Choice on Swag (test)Follow80.85AccuracyClean80.423680.534380.64580.7557May 16, 2023Evaluation ResultsMethodMethodLinksAccuracyUntargeted ScoreTargeted ScoreClean2023.0580.8537.916.74UOR-G2023.0580.5566.1333.87UOR2023.0580.4457.1523.75