List your company

Mathematical Reasoning on Theorem

23.8AccuracyCRPS
6.95211.32615.720.074Mar 20, 2025May 23, 2025Jul 27, 2025Sep 30, 2025Dec 4, 2025Feb 7, 2026Apr 13, 2026

Evaluation results

MethodLinks
CRPSBase Model=LLaMA3-8B, #Samples=60K2026.0423.8
CRPSBase Model=LLaMA3-8B, #Samples=15K2026.0423.5
DARTBase Model=LLaMA3-8B, #Samples=60K2026.0422.9
CRPSBase Model=LLaMA3-8B, #Samples=30K2026.0421.6
SIGMABase Model=LLaMA3-8B, #Samples=60K2026.0421.3
CRPSBase Model=Mistral-7B-v0.1, #Samples=60K2026.0421
MathFusionBase Model=LLaMA3-8B, #Samples=60K2026.0420
CRPSBase Model=Mistral-7B-v0.1, #Samples=30K2026.0419.7
DARTBase Model=LLaMA3-8B, #Samples=590K2026.0419.4
SIGMABase Model=Mistral-7B-v0.1, #Samples=60K2026.0419.2
SIGMABase Model=LLaMA3-8B, #Samples=30K2026.0419.1
DART-Math# Samples=60K, Backbone=Mistral-7B2025.0318.2
DARTBase Model=Mistral-7B-v0.1, #Samples=60K2026.0418.2
MathFusion# Samples=60K, Backbone=Mistral-7B2025.0318.1
MathFusionBase Model=Mistral-7B-v0.1, #Samples=60K2026.0418.1
SIGMABase Model=Mistral-7B-v0.1, #Samples=30K2026.0418
CRPSBase Model=Mistral-7B-v0.1, #Samples=15K2026.0417.8
MathFusion (Sequential)Base Model=LLaMA3-8B, #Samples=30K2026.0417
DARTBase Model=Mistral-7B-v0.1, #Samples=590K2026.0417
WizardMathBase Model=Mistral-7B-v0.1, #Samples=418K2026.0416.6
MMIQCBase Model=LLaMA3-8B, #Samples=2.3M2026.0416.2
RFTBase Model=Mistral-7B-v0.1, #Samples=590K2026.0416.2
MathFusion (Sequential)# Samples=30K, Backbone=Mistral-7B2025.0315.5
MathFusion (Sequential)Base Model=Mistral-7B-v0.1, #Samples=30K2026.0415.5
MathFusion (Parallel)# Samples=30K, Backbone=Mistral-7B2025.0315.2
RFTBase Model=LLaMA3-8B, #Samples=590K2026.0414.9
MetaMathBase Model=Mistral-7B-v0.1, #Samples=400K2026.0414
MathFusion (DeepSeekMath-7B-RL)# Samples=60K, Backbone=DeepSeekMath-7B-RL2025.0313.8
MetaMathBase Model=LLaMA3-8B, #Samples=400K2026.0413.8
MathFusion (Conditional)# Samples=30K, Backbone=Mistral-7B2025.0312.8
MetaMathBase Model=Mistral-7B-v0.1, #Samples=60K2026.0412.2
Mosaic-IT# Samples=15K, Backbone=Mistral-7B2025.039.9
Mosaic-IT + Original GSM8K and MATH# Samples=30K, Backbone=Mistral-7B2025.039.5
GPT Rewritten# Samples=15K, Backbone=Mistral-7B2025.038.9
Standard# Samples=15K, Backbone=Mistral-7B2025.037.6