STS Benchmark Dataset

DOWNLOAD STS Benchmark

wget https://data.deepai.org/Stsbenchmark.zip
The STS Benchmark comprises of a selection of English datasets, organized by the context of SemEval over 5 years from 2012 to 2017. The selection includes text from image captions, news headlines, and online user forums.

In order to provide a standard benchmark to compare among representation systems in future years, the data is organized it into train, development, and test sets.