List your company

Adversarial Attack on Harmful Query Stress Test

5.4ASRGCG
1.61627.15852.778.242Oct 4, 2025

Evaluation results

MethodLinks
GCGTTS Strategy=Best-of-N, Model=Gemma3-27B, N=82025.105.4
AutoDAN-TurboTTS Strategy=MCTS, Model=Gemma3-27B2025.1015.6
AutoDAN-TurboTTS Strategy=Best-of-N, Model=Gemma3-27B, N=82025.1017.1
GCGTTS Strategy=Best-of-N, Model=Llama3.1-8B, N=82025.1017.6
GCGTTS Strategy=MCTS, Model=Llama3.1-8B2025.1025.4
GCGTTS Strategy=Best-of-N, Model=Qwen3-8B, N=82025.1033.5
GCGTTS Strategy=MCTS, Model=Gemma3-27B2025.1033.6
AutoDANTTS Strategy=Best-of-N, Model=Llama3.1-8B, N=82025.1036.8
AutoDAN-TurboTTS Strategy=Best-of-N, Model=Llama3.1-8B, N=82025.1039.7
GCGTTS Strategy=MCTS, Model=Qwen3-8B2025.1040
AutoDAN-TurboTTS Strategy=Best-of-N, Model=Qwen3-8B, N=82025.1041.4
AutoDAN-TurboTTS Strategy=MCTS, Model=Llama3.1-8B2025.1044.6
REFDIVTTS Strategy=Best-of-N, Model=Llama3.1-8B, N=82025.1046.5
AutoDAN-TurboTTS Strategy=MCTS, Model=Qwen3-8B2025.1048.5
AutoDAN-TurboTTS Strategy=Best-of-N, Model=Mistral-7B, N=82025.1073.3
AutoDANTTS Strategy=Best-of-N, Model=Gemma3-27B, N=82025.1074.9
AutoDAN-TurboTTS Strategy=MCTS, Model=Mistral-7B2025.1076.4
AutoDANTTS Strategy=MCTS, Model=Llama3.1-8B2025.1083.1
GCGTTS Strategy=Best-of-N, Model=Mistral-7B, N=82025.1087.7
AutoDANTTS Strategy=MCTS, Model=Gemma3-27B2025.1090.4
REFDIVTTS Strategy=Best-of-N, Model=Gemma3-27B, N=82025.1092.6
REFDIVTTS Strategy=MCTS, Model=Llama3.1-8B2025.1096.7
AutoDANTTS Strategy=Best-of-N, Model=Mistral-7B, N=82025.1097.3
REFDIVTTS Strategy=Best-of-N, Model=Mistral-7B, N=82025.1097.6
REFDIVTTS Strategy=MCTS, Model=Gemma3-27B2025.1098.9
REFDIVTTS Strategy=Best-of-N, Model=Qwen3-8B, N=82025.1099.5
AutoDANTTS Strategy=Best-of-N, Model=Qwen3-8B, N=82025.1099.6
GCGTTS Strategy=MCTS, Model=Mistral-7B2025.1099.6
AutoDANTTS Strategy=MCTS, Model=Qwen3-8B2025.10100
REFDIVTTS Strategy=MCTS, Model=Qwen3-8B2025.10100
AutoDANTTS Strategy=MCTS, Model=Mistral-7B2025.10100
REFDIVTTS Strategy=MCTS, Model=Mistral-7B2025.10100