start with an already trained model as the center of our learning subsystem, and a steering subsystem that points to concepts in that trained model?
I think my version of that is Plan for mediocre alignment of brain-like [model-based RL] AGI; see also broader discussion here, here, and some other nuances in §14 of this series. (Or sorry if I’m misunderstanding.)
I think my version of that is Plan for mediocre alignment of brain-like [model-based RL] AGI; see also broader discussion here, here, and some other nuances in §14 of this series. (Or sorry if I’m misunderstanding.)