I think it does matter for emergent misalignment purposes. I’d think the emergent misalignment is more likely when you teach the model “these things are bad, don’t do them” then fine tune them to do one of those previously prohibited behaviors.
I think it does matter for emergent misalignment purposes. I’d think the emergent misalignment is more likely when you teach the model “these things are bad, don’t do them” then fine tune them to do one of those previously prohibited behaviors.