After staff got back to me about some of my models never getting scanning, I was told the reason these initially failed to scan without admin prodding was the long length of the title with the inclusion of non-English characters in the filename. This would have been really nice to know beyond a private response to a bug report (as far as I'm aware there's no public caution about it anywhere), so I'm letting people know. Some local generation setups will also have issue with "foreign" characters in filename for a LoRA.
The on-site trainer or model page currently lacks any way to actually set a filename for a LoRA. I have however determined a workaround: During training (or possibly during publishing, not yet sure) (It's training. I tested it.) give your model a relatively short (but still informative) name. Then, after the model has been published and scanned, rename it to your desired full (multi-lingual) title. It's a bit ridiculous to require this instead of being able to manually set filename, but this seems to be an effective workaround.