I read Carmilla not that long ago and thought it was kind of interesting, and Uncle Silas long before that, which I was not so into when I read but now think was better than I'd given it credit for, so it was high time to investigate what LeFanu is, I think, more well known for - the short stories*. What was most interesting to me was how few of these stories are actually supernatural in nature, and what's more, how little it matters, even if they are. The trademark of Gothic/fantastic fiction is allegedly the hesitation between rational and supernatural expalantion (says Todorov, but others too), but that is actually a very minor part of these stories. In fact, there's only one story where that hesitation exists and is maintained throughout, but again, it's largely irrelevant, because the issue isn't whether or not the thing tormenting the protagonist is actually a ghost, it's the fact that he's being tormented. The other stories leave it mostly ambiguous whether the other-worldly was involved (though they tend to lean strongly in one direction or the other), but it feels like lip-service.
Meanwhile, the stories are very interested in modern science, and this is a noticeable and interesting feature. There is, for instance, a long digression on how bodies can be identified years after their death, or how certain vapors that link the heart and mind can open the third eye, etc. This makes sense, because the stories are allegedly found among the papers of a doctor, but it gives you a strong sense of a certain kind of world-view which is worth exploring.
A final aspect that is intriguing is precisely this "found manuscript" device - what is called the pseudofactual mode by Barbara Foley and, more recently, Nicholas Paige. This is a pretty late work to still feel the need to deploy that device (though I suspect the pseudofactual pretense lingered a lot longer in Gothic fiction, which is an interesting point), but what is more notable is how ridiculous it is, in terms of how the text works. The stories are meant to be documents found among the doctor's paper by his secretary. Many of them are actually taken from letters the doctor wrote to other people (who were not doctors, which is why there is no scientific jargon, we are told) - but some - like Carmilla are actually reports written by people involved. Others switch narrators. In a lot of the stories, the narrator is basically your standard third person omniscient type, and completely parts ways with the framing of the tale, knowing things the character telling the story couldn't, etc. That's actually a fascinating aspect of the book, and one I intend to explore more.
Overall though, I can't say I strongly recommend this one. LeFanu just doesn't really do it for me, style wise. I'm inclined to say that if you're interested in late 19th century literature, there's a lot of interest going on here, but if you're looking for a spooky story pleasure read, keep it moving.
*Well, that and The House by the Church-yard, which is mostly famous because it was important for Finnegan's Wake. It's free in the amazon Kindle store, so I'll get around to reading it at some point.