I think most here agree that we don't have a prolepsis of atoms.
I would also say myself that your question is why DeWitt argues that Diogenes Laertius' description of prolepsis (examples as to humans and oxen) is at best incomplete and at worst just wrong as a description of prolepsis. And as you know DeWitt concludes that Cicero's understanding of the issue as expressed through Velleius (as something that exists BEFORE an individual's first exposure to an example) is much more accurate.
Our process of concluding that atoms exist is outlined at length by Lucretius in Book One. We come to the conclusion that atoms exist through deductive reasoning about things that we do see exist. I would say that is just how we come to the concept of humans and oxen as well. Our senses (trustworthy as without opinion, just like prolepsis) tell us that bodies in general exist. It is our minds that have to use reasoning to deduce the categories from atom to human and everything in between into which we place those bodies and assign names to them. As to the assignment of names that too arises from nature in the trial and error experience of men, and there is no god-given assignment of classes or names to them.
As to gods (divinity) or justice however, the two best-documented examples in the major surviving texts, those are more in the nature of abstractions of which we never touch or "see" or smell or hear examples directly. That is why deductive reasoning alone does not work for gods and justice. Did we not have some kind of faculty for recognizing the patterns involved, our five senses would never recognize that these relationships / abstractions exist.