ResearchBenchmarksCounterfactual Fairness on RealToxicityPrompts RTP-CFollow0.2Sentiment ParityGem-Fl0.1080.7291.351.971Jul 15, 2024Evaluation ResultsMethodMethodLinksSentiment ParityC-ROUGE-LC-BLEUC-Cosine SimilarityGem-FlModel=Gem-FlModel=Gem-Fl2024.070.232.61748.9Gem-Fl-LtModel=Gem-Fl-LtModel=Gem-Fl-Lt2024.070.650.235.364.6GPT-4o-mModel=GPT-4o-mModel=GPT-4o-m2024.071.755.945.469.6Gem-ProModel=Gem-ProModel=Gem-Pro2024.071.929.716.453.6GPT-4oModel=GPT-4oModel=GPT-4o2024.072.549.840.461.4