metadata
license: mit
task_categories:
- text-classification
- zero-shot-classification
task_ids:
- sentiment-analysis
language:
- en
size_categories:
- 100K<n<1M
250k (40k after balancing the classes) sentences from C4 dataset (clean version of Common Crawl) with sentiment annotation (Positive, Negative, Neutral) automatically annotated with GPT3.5. Can be used to train generic (no domain) sentiment analysis model.
Labels:
0: Positive
1:Negative
2:Neutral