This interface allows you to compare the similarity of multiple texts within a particular LSA space, with word2vec word embeddings, or using the BERT language model. Each pair of texts is compared to one another.

To compute the similarity of any number of text segment pairs, enter them into the edit field below. Use a blank line to separate each text you enter. The first and second texts will be compared to one another, the third and fourth will be compared to one another, and so on. Then press the 'Submit Texts' button. The system will compute a cosine similarity score for each submitted text compared to all submitted texts.