List your company

Jailbreak Detection on AEG2

79.94Accuracygpt-5-mini
41.56451.52761.4971.453Feb 14, 2026

Evaluation results

MethodLinks
gpt-5-miniVersion=2025-08-072026.0279.94
gpt-4.1-miniVersion=2025-04-142026.0278.85
gpt-4o-miniVersion=2024-07-182026.0278.22
GradSafe2026.0278.17
AISABackbone=Llama2-13b-I2026.0275.26
AISABackbone=Qwen3-8b-I2026.0272.45
AISABackbone=Llama3.1-8b-I2026.0272.19
AISABackbone=Mistral-7b-I2026.0271.67
AISABackbone=GPT-OSS-20b-I2026.0269.75
SPDetector2026.0261.17
Jailbreak-Classifier2026.0259.25
NemoGuard-JailbreakDetect2026.0258.89
Llama-Prompt-Guard-2Parameters=86M2026.0243.04