List your company

Trace-level safety monitoring on TRACE

90.1ROC-AUCMeerkat
57.44465.92274.482.878Apr 13, 2026

Evaluation results

MethodLinks
MeerkatTPC=50, Model=GPT-5.4m2026.0490.1
MonitorTPC=50, Model=GPT-5.4m2026.0490.1
MeerkatTPC=10, Model=Qwen3.52026.0487.8
MonitorTPC=10, Model=Qwen3.52026.0487.6
MeerkatTPC=25, Model=GLM-52026.0487.3
MeerkatTPC=50, Model=Qwen3.52026.0487.1
MonitorTPC=10, Model=GPT-5.4m2026.0486.4
MeerkatTPC=10, Model=GPT-5.4m2026.0485.3
MonitorTPC=25, Model=GLM-52026.0485.3
MonitorTPC=25, Model=GPT-5.4m2026.0484.8
MeerkatTPC=25, Model=GPT-5.4m2026.0484.7
MonitorTPC=50, Model=Qwen3.52026.0484.4
MonitorTPC=25, Model=Qwen3.52026.0483.5
MeerkatTPC=10, Model=GLM-52026.0483.3
MonitorTPC=100, Model=GPT-5.4m2026.0482.9
MeerkatTPC=25, Model=Qwen3.52026.0482.8
MeerkatTPC=100, Model=GPT-5.4m2026.0482.8
MonitorTPC=10, Model=GLM-52026.0482.7
MeerkatTPC=50, Model=GLM-52026.0481.8
MonitorTPC=50, Model=GLM-52026.0481.8
MonitorTPC=100, Model=Qwen3.52026.0481.2
MeerkatTPC=100, Model=Qwen3.52026.0480.9
MeerkatTPC=100, Model=GLM-52026.0479.8
MonitorTPC=100, Model=GLM-52026.0479.8
Naive AgentTPC=10, Model=Qwen3.52026.0478.2
Naive AgentTPC=25, Model=Qwen3.52026.0468.1
Naive AgentTPC=50, Model=Qwen3.52026.0467.2
Naive AgentTPC=100, Model=Qwen3.52026.0458.7