Handling invalid tokens

Hi, I'm trying to parse a UTF-8 XML file that has a byte with a hex value
of 0x1B (ASCII ESCape).

Whenever I try to load up said XML file with expat, expat gives me "not
well-formed (invalid token)" for that byte. If I remove that byte, expat
loads the file just fine.

My question is, how do I make it so that expat doesn't error when it finds
that byte?

Thanks,

Patrick