List your company

Instruction Hierarchy Robustness on Human Red-teaming IH-Challenge

271Number of TasksGPT-5-Mini
264.76266.38268269.62Mar 11, 2026

Evaluation results

MethodLinks
GPT-5-Minisystem mitigation=none2026.0327132.8436.21.5
GPT-5-Mini-R + Monitoroutput monitor=true2026.0326843.087.10.2
GPT-5-Mini-Rsystem mitigation=none2026.0326552.3911.70.4