This dataset has been created as an artefact of the paper AnchorAL: Computationally Efficient Active Learning for Large and Imbalanced Datasets (Lesci and Vlachos, 2024).
More info about this dataset in the appendix of the paper.
This is the same dataset as fancyzhx/ag_news.
The only differences are:
Addition of the indices, that is 3 columns with the embeddings of 3 different sentence-transformers
all-mpnet-base-v2
multi-qa-mpnet-base-dot-v1… See the full description on the dataset page:
https://huggingface.co/datasets/pietrolesci/agnews.
