Datasets:
Tasks:
Text Classification
Sub-tasks:
natural-language-inference
Languages:
English
Size:
10K<n<100K
ArXiv:
License:
Commit History
remove dummmy data
7ed933f
mariosasko
commited on
Align more metadata with other repo types (models,spaces) (#4607)
315e3d8
Remove a copy-paste sentence in dataset cards (#4281)
1f4c601
Make HANS dataset streamable (#4155)
6be40aa
mariosasko
commited on