> Is this a gradual property that appears slowly as models scale up? Or something that flips beyond a certain size? I wonder if this also applies to biological

> Is this a gradual property that appears slowly as models scale up? Or something that flips beyond a certain size? I wonder if this also applies to biological neural nets. Some animals seem soo close to humans and yet so far. Also if there could be a way to predict such emergence events. We know that we don't know what could be emergent beyond human intelect, but it would be great to get a quantitative framework about the complexity of the models and emergence.