I'm not too familiar with Expositiones Mathematicae, but have you given them a look?
EDIT: The article I happened to have seen, which made me think that Expo Math might be along the lines Pete Clark was looking for, is this paper of T. Bühler - it modestly claims to no originality save for assembling disparate parts of the literature and writing down what's old news to connoisseurs (I'm paraphrasing here!) but of course this is, in a sense, precisely its originality & worth.
Based on the previous answers, it's not entirely clear how big a difference there is between retaining copyright but transferring an exclusive right to publish, and handing over the copyright. That said, here's a list of places that do, as a standard practice, allow the option of authors retaining copyright.
$\bullet$ Journals published by or affiliated with MSP (including Annals of Mathematics, Algebra and Number Theory, Geometry & Topology, Analysis & PDE, Pacific Journal of Mathematics, etc., a total of 11)
$\bullet$ AMS journals (for most of those the AMS is involved with, but not all)
$\bullet$ Compositio (as the OP noticed)
$\bullet$ SIAM journals (16 journals)
$\bullet$ International Mathematics Research Notices (maybe a few other Oxford journals too). I noticed that I hold the copyright on an article I published there.
$\bullet$ Some of the journals published by the Institute of Mathematical Statistics (like Probability Surveys and Statisical Surveys).
$\bullet$ Some electronic journals, like Integers, and the Electronic Journal of Combinatorics, Theory and Applications of Categories.
$\bullet$ Proceedings of the National Academy of Sciences. Mathematicians occasionally publish here, but it's not a math journal.
Best Answer
The journal Dissertationes Mathematicae publishes quite a lot of entire Ph.D. theses, and the abstracts (which may be fairly long and quite often also include the table of contents of the thesis) are free to view.