List your company

Jailbreak Attack Generation on Jailbreak Prompt Generation

0.23Inference Time (s)APD
-0.22162.82675.8758.9233May 26, 2025

Evaluation results

MethodLinks
APDStudent Model=BERT (109M parameters), Teacher Model=Llama2025.050.2311.68
AutoDAN2025.055.122.29
BlackDAN2025.057.672.67
GCG2025.0511.523.64