IAAR-Shanghai / CRUD_RAG

CRUD-RAG: A Comprehensive Chinese Benchmark for Retrieval-Augmented Generation of Large Language Models
https://arxiv.org/abs/2401.17043
233 stars 20 forks source link

split_merged.json 是怎么生成的?是人工标注的还是基于大模型生成的 #20

Closed pluiest closed 1 month ago

pluiest commented 1 month ago

rt

haruhi-sudo commented 1 month ago

你好,不同数据集生成方式不太一样。有些数据集直接是真实的数据,有些数据集是大模型生成辅以人工校验,可以参考论文中的构建流程。