4 ms·
>Although it's a no no to anthropomorphize on HN, it's worth noting that some folks think humans are post-hoc rationalizers as well: There's enough behavioural
by callmeal 3mo ago
>Although it's a no no to anthropomorphize on HN, it's worth noting that some folks think humans are post-hoc rationalizers as well:
There's enough behavioural research to show that it is the case. For ex:
https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/abs/unconscious-cerebral-initiative-and-the-role-of-conscious-will-in-voluntary-action/D215D2A77F1140CD0D8DA6AB93DA5499 https://www.cambridge.org/core/journals/behavioral-and-brain...
- kfarr 3mo agoThanks, that's a much better example of what I was trying to explain. Again, I'm projecting humanity onto a machine, but it's not a big stretch to imagine that the interaction of attention heads over high-dimensional vectors produces "decisions" that can't be articulated in the moment, only described after the fact in text form.