Abductive Natural Language Inference (aNLI)
ClassificationCommonsenseEnglish
Abductive Natural Language Inference (aNLI) is a classification-focused dataset in English that provides 20 labeled examples distributed in JSON format.
About Abductive Natural Language Inference (aNLI)
Dataset is a binary-classification task, the goal is to pick the most plausible explanatory hypothesis given two observations from narrative contexts. It contains 20k commonsense narrative contexts and 200k explanations."
Details
- Task
- Classification, Commonsense
- Language
- English
- Format
- JSON
- Rows / instances
- 20
- Creator
- Bhagavatula et al.
- Year
- 2019