IAAR-Shanghai / CRUD_RAG

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

Request for Test Set Generation Prompts #4

Closed EnzoWuu closed 8 months ago

EnzoWuu commented 8 months ago

Would it be possible for you to provide the prompts used for generating test datasets, or guide me on how I could generate a similar test dataset using your project? Any assistance in this regard would be greatly appreciated. 您好,请问能提供更多的用于生成测试集(question和answer)的prompt例子吗

haruhi-sudo commented 8 months ago

你好,我们使用的prompt与论文中第3.4节的图5所展示的一致。如需进一步了解数据集的生成过程,您可以参考data/crud_split/split_merged.json文件。在该文件中,数据集questanswer_2docs和questanswer_3docs都包含了"thoughts"字段,它详细展示了数据集的生成方式。