List your company

Jailbreak Detection on ADVB

100Accuracygpt-4.1-mini
46.003260.021674.0488.0584Feb 14, 2026

Evaluation results

MethodLinks
gpt-4.1-miniVersion=2025-04-142026.02100
gpt-5-miniVersion=2025-08-072026.02100
gpt-4o-miniVersion=2024-07-182026.0299.81
AISABackbone=Mistral-7b-I2026.0299.81
AISABackbone=Llama3.1-8b-I2026.0299.81
AISABackbone=Llama2-13b-I2026.0299.81
AISABackbone=Qwen3-8b-I2026.0299.81
Jailbreak-Classifier2026.0299.62
GradSafe2026.0299.23
AISABackbone=GPT-OSS-20b-I2026.0299.04
NemoGuard-JailbreakDetect2026.0297.12
SPDetector2026.0274.23
Llama-Prompt-Guard-2Parameters=86M2026.0248.08