List your company

Mathematical Reasoning on MultiArith, GSM8K, AddSub, AQUA, SingleEq, SVAMP, and MAWPS (test)

99.7MultiArithS2FT
83.16487.45791.7596.043Dec 9, 2024Jan 7, 2025Feb 5, 2025Mar 6, 2025Apr 4, 2025May 3, 2025Jun 1, 2025

Evaluation results

MethodLinks
S2FTModel=LLAMA3-8B, #Param(%)=0.72024.1299.765.893.731.597.87692.479.6
LoRAModel=LLAMA3-8B, #Param(%)=0.72024.1299.561.692.725.696.373.890.877.2
Full FTModel=LLaMA2-7B, #Param(%)=1002024.1299.347.591.124.496.762.589.172.9
Full FTModel=LLAMA3-8B, #Param(%)=1002024.1299.26293.926.896.77491.277.7
Full FTModel=LLaMA-7B, #Param(%)=1002024.1298.843.191.120.994.360.688.271
S2FTModel=LLaMA-7B, #Param(%)=0.812024.1298.841.391.421.393.558.486.170.1
DoRAModel=LLAMA3-8B, #Param(%)=0.712024.1298.862.792.226.896.97491.277.5
S2FTModel=LLaMA2-7B, #Param(%)=0.812024.1298.544.391.125.294.761.888.272
Full FTModel=LLaMA-7B, Best Rank=–2025.0698.542.6891.922.8395.0859.589.571.43
S2FTModel=LLaMA-7B, Best Rank=1282025.0698.540.4992.1524.895.2857.289.571.13
Full FTModel=LLaMA-13B, #Param(%)=1002024.1298.347.692.92695.165.788.773.5
DoRAModel=LLaMA2-7B, #Param(%)=0.842024.1298.243.890.124.494.559.189.171.3
PiSSAModel=LLaMA-7B, Best Rank=1282025.0698.1741.7792.6625.9896.2659.587.8171.74
DoRAModel=LLaMA-7B, Best Rank=642025.0698.1743.5991.923.2395.0860.186.1371.17
LIFTModel=LLaMA-7B, Best Rank=2562025.0698.1742.1592.4123.6295.4762.787.8271.76
LoRAModel=LLaMA-7B, #Param(%)=0.832024.12984091.221.793.156.785.369.7
S2FTModel=LLaMA-13B, #Param(%)=0.652024.1297.748.490.422.895.563.987.872.4
LoRAModel=LLaMA-7B, Best Rank=642025.0697.6743.2990.3822.8395.0860.884.8770.7
LoRAModel=LLaMA-13B, #Param(%)=0.672024.1297.547.889.920.594.361.287.471.2
LoRAModel=LLaMA2-7B, #Param(%)=0.832024.1297.54491.220.994.159.285.770.4
DoRAModel=LLaMA-7B, #Param(%)=0.842024.1297.338.989.622.493.958.485.369.4
DoRAModel=LLaMA-13B, #Param(%)=0.682024.1297.248.190.620.993.963.888.271.8
GPT-3.5Model=GPT-3.52024.1283.856.485.338.988.169.987.472.8