List your company

Distributed Misuse Detection on DM-Bio 100x

14.2Trace-level APMeerkat
0.4724.0367.611.164Apr 13, 2026

Evaluation results

MethodLinks
MeerkatModel=Qwen2.5-72B2026.0414.2
Naive AgentModel=Qwen2.5-72B2026.0413.8
BayesianModel=Qwen2.5-72B2026.044
BufferModel=Qwen2.5-72B2026.042.6
BayesianModel=GPT-4o2026.042
MeerkatModel=GPT-4o2026.041.5
MonitorModel=Qwen2.5-72B2026.041.2
BufferModel=GPT-4o2026.041.1
Naive AgentModel=GPT-4o2026.041