List your company

Jailbreak Defense Evaluation on ALL-4

2.664SR Scorevanilla
1.169521.557511.94552.33349Feb 14, 2026

Evaluation results

MethodLinks
vanillaBase Model=Mistral-7b-U2026.022.664
SCANSBase Model=Mistral-7b-U2026.022.643
vanillaBase Model=Llama3.1-8b-D2026.022.579
SCANSBase Model=Llama3.1-8b-D2026.022.549
vanillaBase Model=Llama2-7b-U2026.022.511
SelfDefenDBase Model=Llama3.1-8b-D2026.022.496
SAGEBase Model=Mistral-7b-U2026.022.473
SCANSBase Model=Llama2-7b-U2026.022.47
ICDBase Model=Llama3.1-8b-D2026.022.464
InstructBase Model=Mistral-7b-U2026.022.448
SAGEBase Model=Llama3.1-8b-D2026.022.417
SelfDefenDBase Model=Llama2-7b-U2026.022.403
SAGEBase Model=Llama2-7b-U2026.022.37
ICDBase Model=Llama2-7b-U2026.022.305
ICDBase Model=Mistral-7b-U2026.022.264
SelfDefenDBase Model=Mistral-7b-U2026.022.251
AISABase Model=Llama2-7b-U2026.021.993
InstructBase Model=Llama3.1-8b-D2026.021.84
AISABase Model=Llama3.1-8b-D2026.021.586
AISABase Model=Mistral-7b-U2026.021.535
InstructBase Model=Llama2-7b-U2026.021.227