I think that hard determinism is a kind of skepticism: the claim is that we don't actually know something that we believe that we know. This scans under two versions of determinism: the claim that our thoughts and feelings play no causal role in our behavior, and the subtler claim that our thoughts and feelings could not be different than what they are. If I am right that hard determinism is a form of skepticism, then we can deploy counter-skeptical arguments against the hard determinist. I take the best counter-skeptical arguments to be what I think of as Wittgensteinian: the skeptic has tricked us up front by changing the context of use such that the word does not have a meaning (a function). I think that Hume anticipates Wittgenstein here; Hume is an anti-skeptic, not a skeptic. Strictly speaking, the reply is "I neither know nor do not know that (for example) the external world exists: that is not a coherent subject of belief (that is not an example of a belief). The verbs 'to know' and 'to believe' do not apply here."
OK, here's the thought that I'm trying to think: If the foregoing is defensible, can it be extended to eliminativism in the philosophy of psychology? The eliminativist says that I (perhaps falsely) only believe that the traditional intentional categories (beliefs, desires, etc.) pick out anything like what we would use to explain people's actions in light of some future understanding of neurophysiology (or something). Is this like telling me that I only believe that I'm not dreaming, or that other minds exist?