Evaluating the Decency and Consistency of Data Validation Tests Generated by LLMs

We investigated the potential of large language models (LLMs) in developing dataset validation tests. We carried out 96 experiments each for both GPT-3.5 and GPT-4, examining different prompt scenarios, learning modes, temperature settings, and roles. The prompt scenarios were: 1) Asking for expectations, 2) Asking for expectations with a given context, 3) Asking for expectations after requesting a simulation, and 4) Asking for expectations with a provided data sample. For learning modes, we tested: 1) zero-shot, 2) one-shot, and 3) few-shot learning. We also tested four temperature settings: 0, 0.4, 0.6, and 1. Furthermore, two distinct roles were considered: 1) "helpful assistant", 2) "expert data scientist". To gauge consistency, every setup was tested five times. The LLM-generated responses were benchmarked against a gold standard suite, created by an experienced data scientist knowledgeable about the data in question. We find there are considerable returns to the use of few-shot learning, and that the more explicit the data setting can be the better. The best LLM configurations complement, rather than substitute, the gold standard results. This study underscores the value LLMs can bring to the data cleaning and preparation stages of the data science workflow.