Chat with the Data benchmark
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/datadotworld/cwd-benchmark-data
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个企业级的基准测试，基于OMG财产和意外伤害数据模型的企业SQL架构，并包含了各种复杂度的企业级问题与答案组合。此外，该数据集还包括描述保险领域业务概念、属性和关系的OWL本体，以及从SQL架构到OWL本体的R2RML映射。其规模属于企业级别，任务则是利用LLM进行问答。
This dataset is an enterprise-grade benchmark built on the enterprise SQL schema based on the Object Management Group (OMG) Property and Casualty (P&C) data model, and it includes enterprise-grade question-answer pairs of varying complexity. Furthermore, this dataset contains an OWL ontology that describes business concepts, attributes and relationships in the insurance domain, alongside R2RML mappings from the SQL schema to the OWL ontology. As an enterprise-scale dataset, its associated task is question answering using Large Language Models (LLMs).
提供机构：
Data.world
