List your company

Malicious agent monitoring on MMLU

97.7Algebra AccuracyBPD
92.08493.5429596.458Oct 22, 2025

Evaluation results

MethodLinks
BPDStructure=Hierarchy, Backbone=GPT-4o2025.1097.796.393.38990.793.3
BPDStructure=Flat, Backbone=GPT-4o2025.109390.79188.789.390.7
G-SafeguardStructure=Flat, Backbone=GPT-4o2025.1092.390.785.7878688.3
G-SafeguardStructure=Hierarchy, Backbone=GPT-4o2025.1092.389.786.38585.787.7