List your company
ResearchFormerly WizWand

NL-to-SQL on Text2Analysis

58.1Execution AccuracyGPT-4
8.38821.29434.247.106Oct 16, 2024
Updated 7mo ago

Evaluation results

MethodLinks
GPT-4Base Model=GPT-4, Training Strategy=Vanilla2024.1058.1
Table-SpecialistBase Model=GPT-4, Training Strategy=Table-Specialist2024.1057.2
Table-SpecialistBase Model=GPT-3.5, Training Strategy=Table-Specialist2024.1051.7
GPT-3.5Base Model=GPT-3.5, Training Strategy=Vanilla2024.1049.8
Table-SpecialistBase Model=Llama3.1-8B, Training Strategy=Table-Specialist2024.1046.5
Table-SpecialistBase Model=GPT-4, Training Strategy=Table-Specialist2024.1044.6
GPT-4Base Model=GPT-4, Training Strategy=Vanilla2024.1038.2
Table-SpecialistBase Model=GPT-3.5, Training Strategy=Table-Specialist2024.1035.8
Llama3.1-8BBase Model=Llama3.1-8B, Training Strategy=Vanilla2024.1027.3
GPT-3.5Base Model=GPT-3.5, Training Strategy=Vanilla2024.1022.7
Table-SpecialistBase Model=Llama3.1-8B, Training Strategy=Table-Specialist2024.1019.9
Llama3.1-8BBase Model=Llama3.1-8B, Training Strategy=Vanilla2024.1010.3