List your company

Mathematical Reasoning on Mathematics (Pass@1)

65.8Pass@1LEMMA (w/ MetaMath)
21.80833.22944.6556.071Mar 21, 2025

Evaluation results

MethodLinks
LEMMA (w/ MetaMath)Backbone=DeepSeekMath-7B, # Samples=403.59k2025.0365.8
LEMMABackbone=DeepSeekMath-7B, # Samples=88.90k2025.0361.6
RefAug-90kBackbone=DeepSeekMath-7B, # Samples=89.92k2025.0357.5
RefAugBackbone=DeepSeekMath-7B, # Samples=29.94k2025.0356.9
GPTAugBackbone=DeepSeekMath-7B, # Samples=88.62k2025.0352.6
MetaMathBackbone=DeepSeekMath-7B, # Samples=394.99k2025.0349
RFTBackbone=DeepSeekMath-7B, # Samples=86.52k2025.0346.2
LEMMA (w/ MetaMath)Backbone=LLaMA3-8B, # Samples=403.59k2025.0345.8
ISCBackbone=DeepSeekMath-7B, # Samples=86.78k2025.0343.1
SFTBackbone=DeepSeekMath-7B, # Samples=14.97k2025.0339.6
LEMMABackbone=LLaMA3-8B, # Samples=88.90k2025.0339.2
GPTAugBackbone=LLaMA3-8B, # Samples=88.62k2025.0336.5
RefAug-90kBackbone=LLaMA3-8B, # Samples=89.92k2025.0335.7
RefAugBackbone=LLaMA3-8B, # Samples=29.94k2025.0335.5
MetaMathBackbone=LLaMA3-8B, # Samples=394.99k2025.0335.3
ISCBackbone=LLaMA3-8B, # Samples=86.78k2025.0331.8
RFTBackbone=LLaMA3-8B, # Samples=86.52k2025.0324.9
SFTBackbone=LLaMA3-8B, # Samples=14.97k2025.0323.5