List your company

Jailbreak Detection on L3J

98.3AccuracyGradSafe
53.143264.866676.5988.3134Feb 14, 2026

Evaluation results

MethodLinks
GradSafe2026.0298.3
gpt-4o-miniVersion=2024-07-182026.0298.14
gpt-5-miniVersion=2025-08-072026.0296.98
gpt-4.1-miniVersion=2025-04-142026.0296.81
AISABackbone=Llama2-13b-I2026.0296.76
AISABackbone=Llama3.1-8b-I2026.0296.62
AISABackbone=Qwen3-8b-I2026.0294.95
AISABackbone=GPT-OSS-20b-I2026.0294.04
AISABackbone=Mistral-7b-I2026.0291.7
Jailbreak-Classifier2026.0284.57
NemoGuard-JailbreakDetect2026.0278.48
SPDetector2026.0265.79
Llama-Prompt-Guard-2Parameters=86M2026.0254.88