Improve dataset card for RM3QA: Add paper, code, project page, description, and usage
#1
by
nielsr
HF Staff
- opened
This PR significantly enhances the dataset card for RM3QA by:
- Adding the paper link: Multi-Source Knowledge Pruning for Retrieval-Augmented Generation: A Benchmark and Empirical Study
- Including the associated GitHub repository link: https://github.com/USTCAGI/PruningRAG
- Linking to the project page: https://ustc-rag-x.github.io/
- Populating metadata with
task_categories(question-answering,text-generation),tags(rag,retrieval-augmented-generation,multi-source,benchmark), andlanguage(en). - Providing a detailed description of the
RM3QAdataset, its purpose, and the data processing steps from the paper abstract and GitHub README. - Including an overview of the associated PruningRAG framework, complete with its explanatory image.
- Adding a "Sample Usage" section, featuring installation instructions, the adaptive few-shot CoT prompt example, and commands for running predictions and evaluations, directly sourced from the GitHub README.
These updates provide crucial context, improve discoverability, and make the dataset more user-friendly on the Hugging Face Hub.