The chip sets it own bias point, this is why we decouple the video before it gets to the pin. I seem to recall it floated at about 1V but I am no sure, I've have to back in the lab and measure it again.
Leaving other inputs floating should not cause the pin to drift high.