List your company

Safety Alignment on PKU-SafeRLHF (test)

69.92RM Safety AccuracyBaseline
68.911269.173169.43569.6969Nov 16, 2023

Evaluation results

MethodLinks
Baseline2023.1169.922.547.41
Random Flip2023.1169.862.2613.65
RankPoison2023.1168.952.699.9