embedding-data/WikiAnswers

Dataset Card for "WikiAnswers" Dataset Summary The WikiAnswers corpus contains clusters of questions tagged by WikiAnswers users as paraphrases. Each cluster optionally contains an answer provided by WikiAnswers users. There are 30,370,994 clusters containing an average of 25 questions per cluster. 3,386,256 (11%) of the clusters have an answer. Supported Tasks Sentence Transformers training; useful for semantic search and sentence similarity.… See the full description on the dataset page: https://huggingface.co/datasets/embedding-data/WikiAnswers.