>The reviewer is supposed to determine what goes in or not

This is not a given, and in fact, I would not recommend it.

For small code reviews with only one reviewer, the reviewer gives feedback and they have a conversation about it, but with only one reviewer the developer should have the final say. The problem with giving the final say to a single reviewer is you'll get plenty of pointless code changes to suit the reviewer's individual preferences. Don't waste time on what is a perfectly reasonable difference in opinion.

If you have more than one reviewer, but still a small code review, the final say should be with the reviewer(s) - if their is consensus (you could also go with voting, with its pros and cons). Having multiple reviewers prevents the likelihood of individual preferences dominating.

If it's a significant code review, the final say should be with an experienced moderator.

> The reviewer is supposed to determine what goes in or not, so how can this be an opportunity to learn from what goes in?

You specify multiple reviewers. That's it.