This is the German translation of the STSbenchmark dataset.
It can be used to train German sentence embeddings like T-Systems-onsite/bert-german-dbmdz-uncased-sentence-stsb.
One translation has been done with deepl.com, the other with Amazon Translate. They can be combined to get some more variation in the data.