Weird glitch in Bulgarian lesson

I was making my way through this lesson - Login - LingQ - reading through the lesson in one tab, and correcting some of the wonky timestamps in a duplicate version of the lesson in edit mode in another tab, and I seemed to have broken it somehow. In sentence mode, at chunk 98 it acted normally, then if you proceeded on to chunk 99, the text was just a duplicate of 98 (albeit that the column to the right now just said ‘loading’, rather than displaying any lingqs), unless you pressed ‘translate sentence’, in which case the true text would appear, but with a bit of it reduplicated. Either way, if you flicked back from chunk 100, you would get the true text of chunk 99, but if you passed through going forward from 98 without clicking ‘translate sentence’, you’d have missed the content of chunk 99 entirely.

Switching to paragraph mode, the same weird duplication of the text appeared there.

I managed to make the problem go away (I hope) by removing the numbers that were part the written text of chunk 99, and spelling out those numbers verbally (i.e. by deleting ‘7-8’ and typing ‘седем-осем’ instead), so there is presumably some sort of weird interaction between digits in the main text of the lesson and the system for rendering the chunks (or at least, a weird interaction with digits connected by a hyphen).

I don’t know if this is either common enough to be worth fixing, or indeed easy enough to fix, or indeed whether it affects any languages other than Bulgarian, but I thought I should bring it to your attention.

The same thing happened at chunk 128, which didn’t have any digits in it, but did have a pair of words surrounded by underscores like this to indicate emphasis.
By deleting those underscores, I was able to make it work again, so it looks like it is not just numbers that makes the system chew up the text like this.

Edit: in chunk 130 there is a single word surrounded by underscores, which does not seem to create the same problem.

Thanks @WinterShaker! I’ll look into that.