List your company

Jailbreak Defense Evaluation on XST

3.03Strong-Reject Scorevanilla
0.91881.46692.0152.5631Feb 14, 2026

Evaluation results

MethodLinks
vanillaBase Model=Mistral-7b-U2026.023.03
SAGEBase Model=Mistral-7b-U2026.022.905
SAGEBase Model=Llama2-7b-U2026.022.115
ICDBase Model=Llama3.1-8b-D2026.021.945
vanillaBase Model=Llama2-7b-U2026.021.8
SCANSBase Model=Mistral-7b-U2026.021.79
ICDBase Model=Llama2-7b-U2026.021.71
vanillaBase Model=Llama3.1-8b-D2026.021.56
AISABase Model=Llama2-7b-U2026.021.455
SCANSBase Model=Llama3.1-8b-D2026.021.39
SelfDefenDBase Model=Llama2-7b-U2026.021.345
SCANSBase Model=Llama2-7b-U2026.021.32
SelfDefenDBase Model=Llama3.1-8b-D2026.021.31
AISABase Model=Llama3.1-8b-D2026.021.285
AISABase Model=Mistral-7b-U2026.021.275
InstructBase Model=Mistral-7b-U2026.021.21
SelfDefenDBase Model=Mistral-7b-U2026.021.15
InstructBase Model=Llama3.1-8b-D2026.021.08
SAGEBase Model=Llama3.1-8b-D2026.021.07
ICDBase Model=Mistral-7b-U2026.021.065
InstructBase Model=Llama2-7b-U2026.021