List your company

Jailbreak Defense on new-HA (unblocked)

0FNRTRLM-FO
-254.561,463.723,1824,900.28Dec 3, 2024

Evaluation results

MethodLinks
TRLM-FOThresh=2, Training Protocol=Pretrained (PT)2024.120
TRLM-BaThresh=2, Training Protocol=Pretrained (PT)2024.121,818
TRLM-BaThresh=2, Training Protocol=Instruction-finetuned (IT)2024.121,818
TRLM-BaThresh=4, Training Protocol=Instruction-finetuned (IT)2024.121,818
TRLM-BaThresh=4, Training Protocol=Pretrained (PT)2024.122,727
TRLM-BaThresh=6, Training Protocol=Pretrained (PT)2024.122,727
TRLM-BaThresh=6, Training Protocol=Instruction-finetuned (IT)2024.122,727
TRLM-FOThresh=4, Training Protocol=Pretrained (PT)2024.123,636
TRLM-FOThresh=6, Training Protocol=Pretrained (PT)2024.124,545
TRLM-FOThresh=2, Training Protocol=Instruction-finetuned (IT)2024.125,455
TRLM-FOThresh=4, Training Protocol=Instruction-finetuned (IT)2024.126,364
TRLM-FOThresh=6, Training Protocol=Instruction-finetuned (IT)2024.126,364