Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
STEM Reasoning on Minerva (Accuracy)
Loading...
32
Accuracy
GeoMin
23.992
26.071
28.15
30.229
Jun 3, 2026
Accuracy
Updated 1mo ago
Evaluation Results
Method
Method
Links
Accuracy
GeoMin
Backbone=Deepseek-R1-D...
2026.06
32
Seq-entropy
Backbone=Deepseek-R1-D...
2026.06
27.6
Self-certainty
Backbone=Deepseek-R1-D...
2026.06
27.6
Tok-entropy
Backbone=Deepseek-R1-D...
2026.06
27.2
TTRL
Backbone=Deepseek-R1-D...
2026.06
25.7
Co-rewarding
Backbone=Deepseek-R1-D...
2026.06
24.6
TraPO
Backbone=Deepseek-R1-D...
2026.06
24.3
Feedback
Search any
task
Search any
task