Depends heavily on the area. I've seen a bunch of bots correcting nonsensical "textbook" issues in languages like Java. The more common the topic among new learners, the more random PRs you get. But apparently even projects like linux kernel aren't safe https://github.com/torvalds/linux/pulls?q=is%3Apr+is%3Aclose... (for context, this repo doesn't accept PRs as development happens in mailing lists)

There are certain companies that “rate” candidates based on e.g. number of open source contributions. Some people then try to game this system by submitting a whole bunch of super minor errors to a lot of repos, e.g. small grammar mistakes / typos.

Wow! That spam seems so bad and dumb that it seems unlikely this would be deterred by the repo being in Japanese.

Why are there so many "Update Reamde.md" pull requests ? They all are named exactly the same, and most of the time add a single line to the readme with the name of the commit author. I guess some git tutorial somewhere shows how to open a PR with this repo as an example, and people actually do it.

In addition to what the other replies have said, just a couple of years ago some company or group was running a competition which gave away something and caused a bunch of nuisance PRs adding almost literal garbage to READMEs. After criticism the ones running the competition eventually said repository owners could opt-out. This is only from recollection so I might have some things wrong about it. Fortunately it hasn't repeated itself.

This sounds like Hacktoberfest with Digital Ocean, specifically. I could also be wrong, but that is the one that stood out to me.