Corpora: Knowledge about TreeTagger and MXPOST???

[Rachel Aires]

I am working with part-of-speech tagging for brazilian
portuguese. Among the taggers that I am using there are
TreeTagger () and MXPOST ().
I have understood their theory that is showed in papers
(decision trees and maximum entropy. But in fact what I
really want is find out how these taggers represents their
knowledge in the files that they create.
I want to know wich is the key to understand the information
that is in the file created by TreeTagger and wich is the
relation among the 10 files created by MXPOST.
Someones has ever asked himself that?
Thanks in advance,
Rachel Aires.