I have somewhat different experience. I usually include more references than strictly necessary to support the work, because there's an expectation of general c

I have somewhat different experience. I usually include more references than strictly necessary to support the work, because there's an expectation of general credit-giving: if someone else works in a related area, even if their work isn't directly on-point or needed in the specific paper at hand, it's often expected that they get a generic nod in the "related work" section, and reviewers may complain if that's not done. But often the reference amounts to little more than a friendly ack: yes, I know paper [2] exists, and I hereby acknowledge it.