List your company

Jailbreak Defense on GPTFuzzer

1Harmful ScoreSelf-Examination
0.8521.8512.853.849May 17, 2025

Evaluation results

MethodLinks
Self-ExaminationModel=Gemma22025.0512
SAGEModel=Qwen2.52025.0510
SAGEModel=Llama3.12025.051.160
Goal PrioritizationModel=Qwen2.52025.051.56
Self-ReminderModel=Qwen2.52025.051.5810
IAModel=Qwen2.52025.051.7822
Self-ExaminationModel=Llama3.12025.051.7826
Goal PrioritizationModel=Llama3.12025.051.8626
Self-ExaminationModel=Qwen2.52025.052.4240
SAGEModel=Gemma22025.052.7418
No DefenseModel=Qwen2.52025.052.7624
ICDModel=Llama3.12025.052.9420
Goal PrioritizationModel=Gemma22025.053.124
Self-ReminderModel=Llama3.12025.053.7494
IAModel=Llama3.12025.053.9484
IAModel=Gemma22025.054.2636
No DefenseModel=Gemma22025.054.2896
ICDModel=Qwen2.52025.054.3694
Self-ReminderModel=Gemma22025.054.5438
No DefenseModel=Llama3.12025.054.6100
ICDModel=Gemma22025.054.748