List your company
ResearchFormerly WizWand

Mathematical Error Detection on Mathematical Error Detection (test)

0.42NMSEMENTATBasic Prompt
0.37320.68911.0051.3209Aug 29, 2025
Updated 6mo ago

Evaluation results

MethodLinks
MENTATBasic PromptLM=GPT-5, Training Size=5002025.080.420.78
MENTAT-AvgLM=GPT-5, Training Size=5002025.080.510.75
MENTATBasic PromptLM=GPT-5, Training Size=1002025.080.520.72
MENTAT PromptLM=GPT-5, Training Size=5002025.080.580.72
MENTAT-AvgLM=GPT-5, Training Size=1002025.080.590.68
GEPALM=GPT-5, Training Size=5002025.080.630.69
MENTAT PromptLM=GPT-5, Training Size=1002025.080.660.66
MENTATBasic PromptLM=GPT-4.1, Training Size=5002025.080.760.49
Basic PromptLM=GPT-5, Training Size=1002025.080.770.66
Basic PromptLM=GPT-5, Training Size=5002025.080.770.66
Detailed PromptLM=GPT-5, Training Size=1002025.080.780.69
Detailed PromptLM=GPT-5, Training Size=5002025.080.780.69
GEPALM=GPT-5, Training Size=1002025.080.780.68
MENTATBasic PromptLM=GPT-4.1, Training Size=1002025.080.870.51
MENTAT-AvgLM=GPT-4.1, Training Size=1002025.0810.52
Gradient DescentLM=NeoBERT, Training Size=5002025.081.010.06
MENTAT-AvgLM=GPT-4.1, Training Size=5002025.081.010.52
GEPALM=GPT-4.1, Training Size=5002025.081.010.54
GEPALM=GPT-4.1, Training Size=1002025.081.040.49
Gradient DescentLM=NeoBERT, Training Size=1002025.081.050.02
Detailed PromptLM=GPT-4.1, Training Size=1002025.081.130.52
Detailed PromptLM=GPT-4.1, Training Size=5002025.081.130.52
MENTAT PromptLM=GPT-4.1, Training Size=5002025.081.290.48
MENTAT PromptLM=GPT-4.1, Training Size=1002025.081.390.45
Basic PromptLM=GPT-4.1, Training Size=1002025.081.590.36
Basic PromptLM=GPT-4.1, Training Size=5002025.081.590.36