Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
LLM Routing on MATH Out-of-Domain
Loading...
73.9
AUROC
ProbeDirichlet
45.924
53.187
60.45
67.713
Feb 12, 2026
AUROC
Updated 4d ago
Evaluation Results
Method
Method
Links
AUROC
ProbeDirichlet
signal_modality=Hidden...
2026.02
73.9
EmbeddingMLP
signal_modality=Embedd...
2026.02
56.97
Entropy
signal_modality=Logit-...
2026.02
55.3
SemanticEntropy
signal_modality=Verbos...
2026.02
55.25
ConfidenceMargin
signal_modality=Logit-...
2026.02
50.05
SelfAsk
signal_modality=Verbos...
2026.02
49.29
MaxLogits
signal_modality=Logit-...
2026.02
47
Feedback
Search any
task
Search any
task