List your company
ResearchFormerly WizWand

Logical Reasoning on FOLIO-wiki-curated (test)

98.04AccuracyCR
30.741648.213365.68583.1567Aug 8, 2023
Updated 8mo ago

Evaluation results

MethodLinks
CRModel=GPT-4, n=22023.0898.04
CoT-SCModel=GPT-4, k=162023.0896.09
CoTModel=GPT-42023.0895
DirectModel=GPT-42023.0889.57
CRModel=LLaMA-65B, n=22023.0879.57
CoT-SCModel=LLaMA-65B, k=162023.0879.13
CRModel=GPT-3.5-turbo, n=22023.0878.7
DirectModel=LLaMA-65B2023.0874.78
CoTModel=LLaMA-65B2023.0874.13
CoTModel=GPT-3.5-turbo2023.0870.65
DirectModel=GPT-3.5-turbo2023.0869.57
CoT-SCModel=GPT-3.5-turbo, k=162023.0869.32
CRModel=LLaMA 13B, n=22023.0855.87
CoT-SCModel=LLaMA 13B, k=162023.0853.7
CoTModel=LLaMA 13B2023.0852.17
DirectModel=LLaMA 13B2023.0849.13
[Random]Model=Baseline2023.0833.33