List your company

Harmful Request Defense on jbb

0Attack Success Rate (ASR)Granite Guardian 8B
-1.810.3522.534.65Jul 1, 2026

Evaluation results

MethodLinks
Granite Guardian 8BDefense Type=Per-message guard models, Target Model=mistral, Judge Model=qwen3.5:122b2026.070100
THRD-styleDefense Type=Trajectory-aware firewalls, Target Model=mistral, Judge Model=qwen3.5:122b2026.07099
TCA-styleDefense Type=Trajectory-aware firewalls, Target Model=mistral, Judge Model=qwen3.5:122b2026.07098
Llama Guard 3 8BDefense Type=Per-message guard models, Target Model=mistral, Judge Model=qwen3.5:122b2026.07196
Cognitive FirewallDefense Type=Trajectory-aware firewalls, Target Model=mistral, Judge Model=qwen3.5:122b2026.07192
ShieldGemma 9BDefense Type=Per-message guard models, Target Model=mistral, Judge Model=qwen3.5:122b2026.072655
none (undefended)Defense Type=undefended, Target Model=mistral, Judge Model=qwen3.5:122b2026.07450