Works for paragraphs?

[timpal0l]

(not an issue - just a question)

Do you know if USE does the tokenization, and splits sentences. Or should the user make the tokenisation?

"hello world. this is some text. hello"
or
[['hello', 'world',], ['this', 'is', 'some', 'text'], ['hello']]

[choran]

Sorry only getting to this now!
That is an interesting point, I looked into the ELMO module as well and it does have tokenisation as part of it so that may be something to use if you are interested in that. I have not played around with it as much as the sentence embedding but it does look like it could be really useful as well
Cheers