This dataset is a collection of question-answer-negative triplets and question-answer-negative_1...-negative_5 tuples from the Natural Questions dataset. See Natural Questions for additional information.
This dataset can be used directly with Sentence Transformers to train embedding models.
The negative samples have been automatically mined with all-MiniLM-L6-v2 and:
range_min: 10, i.e. we skip the 10 most similar samples… See the full description on the dataset page:
https://huggingface.co/datasets/tomaarsen/natural-questions-hard-negatives.
