I think I expect all of those to be more or less accurate (for the >= 10 ones).
If someone says a dozen, I would expect exactly 12 (especially if it’s something with very strong collocation like “eggs”). If someone says “a baker’s dozen” I would expect exactly 13. If someone says “score”, I’d be surprised, but given how infrequently it’s used, I’d expect the speaker to be using it accurately, not sloppily.