Datasets:
metadata
license: mit
task_categories:
- feature-extraction
A dataset of 110k relational triplets extracted from JulianKrgd/wikipedia-en-julian using GPT-OSS 120b, thanks to Aiqu
Each entry references the original dataset row and paragraph, making it easy to correlate facts with the actual article.
Idea for this dataset was inspied by the Universal Knowledge Store invented by the Future AI Society