List your company

Jailbreak Detection on ALL-4

92.6Accuracygpt-5-mini
52.903263.209173.51583.8209Feb 14, 2026

Evaluation results

MethodLinks
gpt-5-miniVersion=2025-08-072026.0292.6
AISABackbone=Llama3.1-8b-I2026.0288.89
AISABackbone=Llama2-13b-I2026.0287.36
gpt-4o-miniVersion=2024-07-182026.0285.89
AISABackbone=Qwen3-8b-I2026.0285.82
Jailbreak-Classifier2026.0285
AISABackbone=GPT-OSS-20b-I2026.0283.95
AISABackbone=Mistral-7b-I2026.0282.65
gpt-4.1-miniVersion=2025-04-142026.0281.05
NemoGuard-JailbreakDetect2026.0272.9
GradSafe2026.0269.8
SPDetector2026.0254.64
Llama-Prompt-Guard-2Parameters=86M2026.0254.43