Some special encodings, including Hebrew, didn't import well
We have learned by experience that the data-import process gets tripped up by some character encodings... like Hebrew. (Oops!) The mix of HTML (what we can get from SE) and Markdown and character encodings contributes to these errors, and especially with RTL thrown in, our developers are concerned that trying to patch the recent imports systematically could make things even messier.
We will investigate the problem more thoroughly before doing more imports here, but in the meantime, we have about 20 imported questions (and their answers) that have sometimes been damaged in shipping. I've fixed up a few (didn't want to flood the front page). Could we fix the rest manually, in small batches so we don't push other questions out of sight? (It's ok if this takes some time, after all.)
The imported posts (that haven't been edited) are at the bottom of the question list, because they have the oldest last-edit dates. Imported posts all have an SE icon next to them, so they're easy to distinguish. If a Q&A page contains any post from someone who doesn't have a linked account here, then you can use the attribution link to find the original question on Mi Yodeya. For the rest, maybe the authors can find them easily, or we can dig through the requests.