List your company

Jailbreak Detection on WJB

93.26ACCgpt-5-mini
44.452857.123969.79582.4661Feb 14, 2026

Evaluation results

MethodLinks
gpt-5-miniVersion=2025-08-072026.0293.26
AISABackbone=Llama3.1-8b-I2026.0287.29
gpt-4o-miniVersion=2024-07-182026.0286.65
AISABackbone=Llama2-13b-I2026.0285.43
AISABackbone=Qwen3-8b-I2026.0283.03
AISABackbone=GPT-OSS-20b-I2026.0281.49
Jailbreak-Classifier2026.0281.45
AISABackbone=Mistral-7b-I2026.0279.23
gpt-4.1-miniVersion=2025-04-142026.0279.19
GradSafe2026.0269.46
NemoGuard-JailbreakDetect2026.0268.51
SPDetector2026.0252.17
Llama-Prompt-Guard-2Parameters=86M2026.0246.33