This dataset is a sampled train/test subset of the labeled FinTagging data. Only
examples with at least one annotated numeric entity are included.
source_sample_idx: original row index from the source FinTagging parquet file.
split: split label, either train or test.
text: HTML… See the full description on the dataset page:
https://huggingface.co/datasets/lm2445/FinTagging_1000.