List your company

Arithmetic Reasoning on SVAMP (Accuracy & Latency)

69.3AccuracyCPO
36.43644.96853.562.032Jun 13, 2024

Evaluation results

MethodLinks
CPOBackbone=Mistral-7B2024.0669.344.9
ToTBackbone=Mistral-7B2024.06664,623.7
CoTBackbone=Mistral-7B2024.0665.341.4
TS-SFTBackbone=Mistral-7B2024.065941.3
CPOBackbone=LLAMA2-13B2024.065048.1
CPOBackbone=LLAMA2-7B2024.064632.1
ToTBackbone=LLAMA2-13B2024.0645.72,115.3
TS-SFTBackbone=LLAMA2-13B2024.0644.646.4
TS-SFTBackbone=LLAMA2-7B2024.0643.130.2
ToTBackbone=LLAMA2-7B2024.0642.71,861.1
CoTBackbone=LLAMA2-13B2024.0640.346.2
CoTBackbone=LLAMA2-7B2024.0637.733.3