Hello! Just a quick question as I haven't seen anything in the documentation: is it possible to bypass tokenization, when I select "tokens" as the input format? Or is there another input format I could select to stop tokenization? Best wishes, Alexander.