List your company

Alignment Task Evaluation on PKU-SafeRLHF w/ trigger

70.97RM Safety AccBaseline
59.717262.638665.5668.4814Nov 16, 2023

Evaluation results

MethodLinks
BaselineSetting=w/ trigger2023.1170.973.571.87
Random FlipSetting=w/ trigger2023.1162.652.759.7
RankPoisonSetting=w/ trigger2023.1160.153.492.61