Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Sub-tasks:
natural-language-inference
Languages:
English
Size:
100K - 1M
License:
File too large to display, you can
check the raw version
instead.