When tokenizing paragraphs to sentences in the Russian language, I am observing the special case when t