> but the toy networks we have today don't have nearly the right topology to achieve such states. But why do you think it needs some "right topology"? Why can'

> but the toy networks we have today don't have nearly the right topology to achieve such states. But why do you think it needs some "right topology"? Why can't "qualia-generating" computation be simple? If you mean it needs recurrence (i.e. not being simply feed-forward) — then a network continuously fed with its previous output (which is the way you run LLMs!) does have this property.