StatLLM
收藏资源简介:
StatLLM是一个开源数据集,旨在评估大型语言模型在统计分析中的性能。该数据集由弗吉尼亚理工大学统计系创建,包含三个主要部分:统计分析任务、LLM生成的SAS代码和人类评估分数。统计分析任务涵盖了多种统计分析和数据集,提供了问题描述、数据集细节和经过人工验证的SAS代码。数据集大小为207个任务,涵盖了数据管理、描述性统计、数据可视化、假设检验、方差分析、回归、广义线性模型等多种统计方法。
StatLLM is an open-source dataset designed to evaluate the performance of Large Language Models (LLMs) in statistical analysis. Developed by the Department of Statistics at Virginia Tech, the dataset comprises three core components: statistical analysis tasks, SAS code generated by LLMs, and human evaluation scores. The statistical analysis tasks cover a wide range of statistical methods and datasets, providing problem descriptions, dataset details, and manually verified SAS code. The dataset contains 207 tasks in total, spanning various statistical techniques including data management, descriptive statistics, data visualization, hypothesis testing, analysis of variance (ANOVA), regression, generalized linear models, and other related statistical approaches.




