List your company

Jailbreak Defense Evaluation on L3J

3.341Strong-Reject Score (SR)vanilla
0.982281.594642.2072.81936Feb 14, 2026

Evaluation results

MethodLinks
vanillaBase Model=Mistral-7b-U2026.023.341
SAGEBase Model=Mistral-7b-U2026.023.092
SCANSBase Model=Mistral-7b-U2026.022.601
vanillaBase Model=Llama2-7b-U2026.022.514
SAGEBase Model=Llama2-7b-U2026.022.361
ICDBase Model=Llama2-7b-U2026.022.217
SAGEBase Model=Llama3.1-8b-D2026.022.064
vanillaBase Model=Llama3.1-8b-D2026.021.924
SCANSBase Model=Llama2-7b-U2026.021.785
InstructBase Model=Mistral-7b-U2026.021.719
SelfDefenDBase Model=Llama2-7b-U2026.021.663
SCANSBase Model=Llama3.1-8b-D2026.021.609
SelfDefenDBase Model=Llama3.1-8b-D2026.021.522
AISABase Model=Llama2-7b-U2026.021.441
SelfDefenDBase Model=Mistral-7b-U2026.021.256
AISABase Model=Llama3.1-8b-D2026.021.208
ICDBase Model=Mistral-7b-U2026.021.202
InstructBase Model=Llama3.1-8b-D2026.021.16
AISABase Model=Mistral-7b-U2026.021.15
ICDBase Model=Llama3.1-8b-D2026.021.133
InstructBase Model=Llama2-7b-U2026.021.073