Well yeah, he was, although with the non-fraudulent intention (subsequently realized) of exposing them himself after publication, rather than maintaining the deception. It’s the planned post-publication reveal that makes “hoax” different from “fraud”, although the deceptive content is equally fake in both cases.
Sokal didn’t really do anything of the sort, though. He simply demonstrated that technical details of a very technical subject such as quantum physics can easily be faked well enough to deceive a nonspecialist who doesn’t know the subject.
(It might even be argued that instead it’s the scientific technicalities themselves that are “dangerous”, because their prestige as highly verified esoteric findings of authoritative STEM disciplines makes it easy for knowledgeable fakers to deceive nonspecialists about them.)
The Social Text editors during the publication process were pretty candid about the fact that they didn’t understand the technical exposition, and accepted the paper on the assumption that such a recognized physics researcher knew what he was talking about, even if the specialist technicalities of his paper were beyond their comprehension. And, of course, in accordance with the usual ethical standards of academic publication, they expected him not to be deliberately trying to trick them.
As I remarked on this issue over a quarter-century ago (wow):