18-12-2020, 10:42 AM
I was always intrigued by the CLS.
I am curious about the "exception" words, which if I remember correctly make up about 30%, most of the time differing by only one token.
I wonder how these exceptions might relate to effects described by Emma and Marco about positional transformations of words at the beginning or end of lines.
Is there a correlation between the exception words and their position?
I am curious about the "exception" words, which if I remember correctly make up about 30%, most of the time differing by only one token.
I wonder how these exceptions might relate to effects described by Emma and Marco about positional transformations of words at the beginning or end of lines.
Is there a correlation between the exception words and their position?