The training set is huge and model "forgets" some of the stuff, providing docs in context makes sense, plus docs could be up to date comparing to training set.
I'd say pull request should be reverted if necessary, but an idea that each commit should be revertable and expectation that project should work after it is unneeded complexity for me.
Sorry for off-topic, just wanted to say that the price page on mullvad.net is the best one I saw comparing to other SaaS. Everything simple and straight forward!