I have to do this a fair bit. My background is PHP, but as a lead I often find myself reviewing JavaScript, and config files for CI/CD that I don't entirely understand. While I understand the basic syntax, there are a lot of functions and patterns that are alien to me.
So I focus most of my attention on test cases. Do they cover all the edge cases we discussed in sprint planning? Now that the implementation is done, can I see any more that area necessary? How clear are the test cases?
Then I turn to the general readability of the code. Are there any strange variable names? Any complicated blocks that could be made clearer? Any strange bits of magic that could have a comment? I might also look for other common anti-patterns - duplicated code, magic numbers, dead code.
Some of this depends on the idiomatic use of the language, and the code style the team has agreed on. There will be places where you'll be wrong, and the developers will be doing the right thing, despite it looking strange to you.
But pushing people to explain their decisions can make them better developers, as they'll have to think about why they code in that way instead of "I dunno, I just did".