Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
LLM Routing on MMLU Pro Humanities Out-of-Domain
Loading...
51.74
LPM
ProbeDirichlet
49.3272
49.9536
50.58
51.2064
Feb 12, 2026
LPM
MPM
HCR
Updated 4d ago
Evaluation Results
Method
Method
Links
LPM
MPM
HCR
ProbeDirichlet
Signal Modality=Hidden...
2026.02
51.74
64.51
11.5
MaxLogits
Signal Modality=Logit-...
2026.02
51.19
62.15
9.25
Entropy
Signal Modality=Logit-...
2026.02
50.77
63.01
9.23
SelfAsk
Signal Modality=Verbos...
2026.02
50.58
62.94
10.75
EmbeddingMLP
Signal Modality=Embedd...
2026.02
50.26
62.53
9.25
ConfidenceMargin
Signal Modality=Logit-...
2026.02
49.5
61.87
12.25
SemanticEntropy
Signal Modality=Verbos...
2026.02
49.42
61.64
10.75
Feedback
Search any
task
Search any
task