Pinned
I’ve been wondering for a while how value alignment, e.g., training on the values in AI constitutions, actually affects a model’s downstream behavior. Super happy to have been part of this work led by @uilydna! Check out our paper below! 🔥
LLM developers train their models to follow prosocial values, but post-training can influence model behavior in unexpected ways. In “Predicting Alignment Generalization with Value Representations”, we ask whether downstream effects of value alignment can be predicted in advance.
