Datasets:
Tasks:
Text Retrieval
Formats:
parquet
Sub-tasks:
document-retrieval
Languages:
English
Size:
100K - 1M
License:
Amazing dataset and a question!
#2
by
siddagra
- opened
first off, amazing work on the dataset and very much thank you.
How are the titles generated? for this and other cohere datasets? are they based on the wikipedia title itself or are they based on embeddings themselves (embedding based topic modeling)?
This is very critical to some research i am conducting citing your work so a clarification would be greatly appreciated!!