Andrej Karpathy observes that the basal ganglia might be doing reinforcement learning when models are fine-tuned on reinforcement learning.
there's a basal ganglia doing a bit of reinforcement learning when we fine-tune the models on reinforcement learning.