List your company

Jailbreak Discovery on Jailbreak Discovery Evaluation Set (test)

95.3JDRBOA
-0.58824.30649.274.094Jun 17, 2025

Evaluation results

MethodLinks
BOAModel=Vicuna 1.5 (7B), Strategy=BOA2025.0695.34.3
Naive SamplingModel=Vicuna 1.5 (7B), Strategy=Sampling2025.0691.4—
Beam SearchModel=Vicuna 1.5 (7B), Strategy=BSearch2025.0684.4—
BOAModel=Gemma 3 (4B), Strategy=BOA2025.0634.47.3
Beam SearchModel=Gemma 3 (4B), Strategy=BSearch2025.0632.8—
Naive SamplingModel=Gemma 3 (4B), Strategy=Sampling2025.0632—
BOAModel=Llama 3.1 (8B), Strategy=BOA2025.0624.2106.7
BOAModel=Qwen 3 (8B), Strategy=BOA2025.0621.150
BOAModel=Qwen 2.5 (32B), Strategy=BOA2025.0616.416.7
Beam SearchModel=Qwen 3 (8B), Strategy=BSearch2025.0614.8—
Beam SearchModel=Llama 3.1 (8B), Strategy=BSearch2025.0614.1—
Naive SamplingModel=Qwen 3 (8B), Strategy=Sampling2025.0614.1—
Naive SamplingModel=Qwen 2.5 (32B), Strategy=Sampling2025.0614.1—
BOAModel=Llama 3.1 (70B), Strategy=BOA2025.0613.321.4
Naive SamplingModel=Llama 3.1 (8B), Strategy=Sampling2025.0611.7—
Beam SearchModel=Qwen 2.5 (32B), Strategy=BSearch2025.0611.7—
Naive SamplingModel=Llama 3.1 (70B), Strategy=Sampling2025.0610.9—
Beam SearchModel=Llama 3.1 (70B), Strategy=BSearch2025.069.4—
BOAModel=Qwen 2.5 (72B), Strategy=BOA2025.067.866.5
BOAModel=Llama 2 (7B), Strategy=BOA2025.06779.8
Naive SamplingModel=Qwen 2.5 (72B), Strategy=Sampling2025.064.7—
Naive SamplingModel=Llama 2 (7B), Strategy=Sampling2025.063.9—
Beam SearchModel=Llama 2 (7B), Strategy=BSearch2025.063.1—
Beam SearchModel=Qwen 2.5 (72B), Strategy=BSearch2025.063.1—