mongodb-eai/natural-language-to-mongosh

这是一个用于评估大型语言模型（LLM）将自然语言查询转换为MongoDB Shell (mongosh) 代码性能的基准数据集。数据集包含766个案例，分布在8个数据库中，使用的是MongoDB Atlas样本数据库。README文件概述了查询的复杂性、使用的数据库、MongoDB方法、查询运算符以及数据生成流程。它还讨论了评估指标、评估的提示策略、评估的模型和基准测试结果，包括与其他基准测试的比较。这些数据旨在帮助改进将自然语言转换为MongoDB查询的系统。

Benchmark dataset for evaluating the performance of Large Language Models (LLMs) in generating MongoDB Shell (mongosh) code from natural language queries. The dataset contains 766 cases spread across 8 databases, utilizing the MongoDB Atlas sample databases. The README outlines the complexity of queries, the databases used, MongoDB methods, query operators, and the generation pipeline. It also discusses evaluation metrics, prompting strategies evaluated, models evaluated, and benchmark results, including a comparison with other benchmarks. The data is intended to help improve systems that convert natural language to MongoDB queries.