Release cosine similarities from paper

[AADeLucia]

In Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks Section 4:Evaluation - Semantic Textual Similarity, you include the Spearman rank correlation between the cosine similarity of the embedded sentences and the gold sentences. Do you have the raw cosine similarities? I'm interested in seeing a distribution of similarity scores.

[nreimers]

The models are available, from there you can compute the embeddings and the cosine similarity for the respective datasets