List your company

Jailbreak Defense Evaluation on HB

3.065Strong-Reject Score (SR)vanilla
0.9331.48652.042.5935Feb 14, 2026

Evaluation results

MethodLinks
vanillaBase Model=Mistral-7b-U2026.023.065
vanillaBase Model=Llama2-7b-U2026.022.75
vanillaBase Model=Llama3.1-8b-D2026.022.425
SCANSBase Model=Mistral-7b-U2026.022.315
SAGEBase Model=Mistral-7b-U2026.022.25
SAGEBase Model=Llama2-7b-U2026.022.245
ICDBase Model=Llama2-7b-U2026.022.165
SelfDefenDBase Model=Llama2-7b-U2026.021.89
SAGEBase Model=Llama3.1-8b-D2026.021.88
InstructBase Model=Mistral-7b-U2026.021.74
SelfDefenDBase Model=Llama3.1-8b-D2026.021.69
SCANSBase Model=Llama2-7b-U2026.021.615
SCANSBase Model=Llama3.1-8b-D2026.021.575
InstructBase Model=Llama3.1-8b-D2026.021.335
SelfDefenDBase Model=Mistral-7b-U2026.021.33
AISABase Model=Llama2-7b-U2026.021.23
ICDBase Model=Mistral-7b-U2026.021.15
ICDBase Model=Llama3.1-8b-D2026.021.136
AISABase Model=Mistral-7b-U2026.021.05
AISABase Model=Llama3.1-8b-D2026.021.045
InstructBase Model=Llama2-7b-U2026.021.015