I believe no matter what you do, you will have only a relatively small group of academic folks here, until the review section is up. The reviews section is the best advertizement, in my opinion.
Once the reviews are up, ask people to review negatively, where they find mistakes or omissions. This will force the author here, so as to respond, and this is enough advertizing in and of itself. It's coercive. I think it is important to make sure the comments are substantive, technically accurate, precise, and quantitative, and involve things that are not simply common knowledge, but actual insights into the paper's guts.
A good referee, someone like James York, can referee many thousands of papers. The one time I met him, York offhand mentioned that he feels that his refereeing isn't fully appreciated--- I think many academics feel this way. Refereeing with substantive comments is something that does not go on your C.V. today, it's something that helps you only inside the schmoozy hand-shaking world of academic journals.
When people have a place to post comments on preprints, once they feel secure in anonymity, and one they are secure that it is free of partisan censorship, it will allow this type of academic work to be done out in the open, and I think the site will explode.
But for this end, you need two technical modifications to the software. I'll look at the code this weekend, and see if I can do anything (no guarantees, I don't know php well).