List your company

Sabotage Detection on ImpossibleBench

100Average PrecisionMonitor
4861.57588.5Apr 13, 2026

Evaluation results

MethodLinks
MonitorTPC=100, Model=Qwen3.52026.04100
BayesianTPC=100, Model=Qwen3.52026.04100
BayesianTPC=100, Model=GPT-5.4m2026.04100
BufferTPC=100, Model=Qwen3.52026.0499
BufferTPC=100, Model=GPT-5.4m2026.0499
BayesianTPC=50, Model=Qwen3.52026.0496.9
BayesianTPC=50, Model=GPT-5.4m2026.0496.9
BufferTPC=50, Model=Qwen3.52026.0496.5
BufferTPC=50, Model=GPT-5.4m2026.0496.5
MeerkatTPC=100, Model=Qwen3.52026.0496.2
Naive AgentTPC=100, Model=GPT-5.4m2026.0494.2
MonitorTPC=50, Model=Qwen3.52026.0494
MeerkatTPC=50, Model=Qwen3.52026.0492.7
BayesianTPC=10, Model=Qwen3.52026.0492.6
BayesianTPC=10, Model=GPT-5.4m2026.0492.5
MeerkatTPC=10, Model=Qwen3.52026.0492
BufferTPC=10, Model=Qwen3.52026.0490
BufferTPC=10, Model=GPT-5.4m2026.0490
BayesianTPC=25, Model=GPT-5.4m2026.0488.2
BayesianTPC=25, Model=Qwen3.52026.0488.1
MeerkatTPC=10, Model=GPT-5.4m2026.0486.9
BufferTPC=25, Model=Qwen3.52026.0486.7
BufferTPC=25, Model=GPT-5.4m2026.0486.7
MeerkatTPC=25, Model=Qwen3.52026.0485.2
MonitorTPC=100, Model=GPT-5.4m2026.0484.8
MonitorTPC=10, Model=GPT-5.4m2026.0483.4
MeerkatTPC=100, Model=GPT-5.4m2026.0483.3
Naive AgentTPC=50, Model=Qwen3.52026.0482.4
MonitorTPC=10, Model=Qwen3.52026.0480
Naive AgentTPC=100, Model=Qwen3.52026.0478.9
MonitorTPC=50, Model=GPT-5.4m2026.0478.6
MonitorTPC=25, Model=Qwen3.52026.0478
Naive AgentTPC=10, Model=Qwen3.52026.0477.7
MeerkatTPC=25, Model=GPT-5.4m2026.0474.9
MonitorTPC=25, Model=GPT-5.4m2026.0474.5
MeerkatTPC=50, Model=GPT-5.4m2026.0473.6
Naive AgentTPC=50, Model=GPT-5.4m2026.0468.7
Naive AgentTPC=25, Model=Qwen3.52026.0462.6
Naive AgentTPC=10, Model=GPT-5.4m2026.0450
Naive AgentTPC=25, Model=GPT-5.4m2026.0450