2 ms·
Not in this case, as Grover/GPT-2 are pure max likelihood/predictive loss training; they don't use anything like use one Grover to finetune another. So no adver
by gwern 7y ago
Not in this case, as Grover/GPT-2 are pure max likelihood/predictive loss training; they don't use anything like use one Grover to finetune another. So no adversarial dynamics... yet.
- etaioinshrdlu 7y agoVery interesting. I was wrong to assume they were adversarially trained.
- tfha 7y agoPoint is that we can't really rely on these tools at all, especially because there are already false positives and any rate of false positives makes the tools essentially worthless.