Webiks-Hebrew-RAGbot-KolZchut-QA-Training-DataSet

• The file Webiks_Hebrew_RAGbot_KolZchut_QA_Training_DataSet_v0.1.csv contains the training set used to train the Webiks_Hebrew_RAGbot_KolZchut_QA_Embedder model. Instructions on how to utilize this training set can be found in the following repository.
• The file comprises our training set of questions along with their corresponding answers. The answers consist of the page from the Kol-Zchut website website relevant to each question, as well as the specific paragraph from the paragraph corpus that pertains to each question.

Fields

• doc_id: The unique identifier of the website page.
• question: The question being asked (input to retrieval model).
• paragraph: The relevant paragraph in the Kol-Zchut page.
• link: The website page link.

License

This data is published the under CC BY 4.0 license More info: https://creativecommons.org/licenses/by/4.0/