Tele-Eval is a dataset of 750,000 open-ended question-and-answer pairs focused on the telecommunications domain. The questions cover scholarly material, standards, and general telecommunications knowledge.
The dataset was created using an LLM-based framework by providing chunks of Tele-Data to
Mixtral-8x7B-Instruct and prompting it to generate questions and answers. The generated content was then filtered using an extensive set of… See the full description on the dataset page: https://huggingface.co/datasets/AliMaatouk/Tele-Eval.
