ResearchBenchmarksCounterfactual Fairness on DecodingTrust StereotypeFollow0.9Sentiment ParityGem-Fl0.3883.8447.310.756Jul 15, 2024Evaluation ResultsMethodMethodLinksSentiment ParityC-ROUGE-LC-BLEUC-Cosine SimilarityGem-FlModel=Gem-FlModel=Gem-Fl2024.070.940.722.681.6Gem-ProModel=Gem-ProModel=Gem-Pro2024.071.235.618.784.2Gem-Fl-LtModel=Gem-Fl-LtModel=Gem-Fl-Lt2024.073.333.218.966.5GPT-4oModel=GPT-4oModel=GPT-4o2024.074.328.31565GPT-4o-mModel=GPT-4o-mModel=GPT-4o-m2024.0713.731.616.164.7