Few journals and conferences so far seem to use open review. We mostly see open review practised as post-publication commenting on for example pubpeer.com where it so far seems to be mainly about spotting errors in already published papers.
I would personally like to see more open review employed by journals and conferences in the publishing of scientific papers to increase transparency in the process.
Today I have found such an example thanks to Igor Carron’s post The papers for ICLR 2015 are now open for discussion! The machine learning conference International Conference on Learning Representations uses an open review model where reviews are published, anyone can comment on the papers, and anyone can ask to become a designated reviewer: http://www.iclr.cc/doku.php?id=pubmodel.
Even though independent sites exist for post-publication commenting and review, I think it is especially exciting to see it being actively encouraged and fully integrated into the paper submission and acceptance process by the conference organisers. In addition to providing transparency in the process, I hope it also stimulates more discussion when the it is actively encouraged as we see here.