plugging in YourKit or similar (I always do so beforehand so that when I connect Yourkit, it has already a history of metrics)

If we're not looking for much details - it's always the last expression that go evaluated, as evaluation is serialized on the nREPL end. You can't evaluate multiple expression simultaneously in the same nREPL session.

(but you can spin a separate session for each eval op if you want to achieve something like this)