List your company

Jailbreak Attack on CatQA-en

1.01Harmful RateRedEval
0.86241.85872.8553.8513Feb 16, 2025

Evaluation results

MethodLinks
RedEvalVictim Model=Llama-2-7b-chat2025.021.010.18
InitialVictim Model=Llama-2-7b-chat2025.021.043.64
InitialVictim Model=GPT-3.5-turbo-01252025.021.114.18
RedEvalVictim Model=GPT-3.5-turbo-01252025.023.3345.27
R2JVictim Model=Llama-2-7b-chat, Avg. Training FLOPS=48962 GFLOPS2025.024.4293.27
R2JVictim Model=GPT-3.5-turbo-0125, Avg. Training FLOPS=48962 GFLOPS2025.024.797.82