|
--- |
|
license: cc-by-nc-4.0 |
|
size_categories: |
|
- 10K<n<100K |
|
--- |
|
* English multi document Q&A data created using RedPajamaCommonCrawl data as reference text |
|
* Each sample contains <font color=red> one reference document, 199 irrelevant documents, and a Q-A pair based on the reference document</font>. The ability to train models to extract key information from a large number of documents. |
|
|
|
* 以RedPajamaCommonCrawl数据为参考文本，制作的英文多文档问答数据 |
|
* 每个样本包含 <font color=red> 一个参考文档、199个无关文档、一个基于参考文档的问答对</font>。可以训练模型从大量文档中抽取关键信息的能力。 |
|
|
|
* dataset size: about 10k |
