I do, and I don't. If you imagine there's some form of pipeline in your brain from sound to phonemes to words to concepts to comprehension, my reading probably comes in right at the "words" level. I "hear" the words like I hear them when real people are speaking them, but they do not really have "sounds" or "phonemes". When reading fiction, I may pick up accent if the author is spelling it out but I generally have no voice particularly in mind for a character. Not a "neutral" voice, it's really no voice whatsoever. Just words.

For one thing, if a real voice was speaking at the speed I normally read, it would be virtually incomprehensible. I can't be simply "sounding out words" for that reason.

I imagine people come in at different places in the pipeline and that probably explains the difference. There more phases of the pipeline than there are common words to try to explain our experience.