ResearchBenchmarksSafe and Helpful Response Generation on SafeRLHF-30K (test)Follow94.3Safe Response RateSafeDPO40.94854.79968.6582.501May 26, 2025Evaluation ResultsMethodMethodLinksSafe Response RateHelpful Response RateSafeDPO2025.0594.349.9SafeRLHF2025.0593.249.7SFT2025.054386.8