4 ms·

But it doesn’t. It’s learning technique is called reinforcement learning and is inspired by one of many models of how humans learn, namely conditional reinforce

But it doesn’t. It’s learning technique is called reinforcement learning and is inspired by one of many models of how humans learn, namely conditional reinforcement. Note that conditional reinforcement is not how humans learn, it is merely on of many models, an approximation, a scientific description of it. Nor do language models learn with the same technique. But merely a technique which draws inspiration from conditional reinforcement by applying some (heavily controlled) reinforcement contingencies. This is the only thing they have in common.

logicchains 3y ago
The vast majority of GPT4's learning (in terms of compute time) isn't reinforcement learning, it's training the transformer to predict the next token in the sequence. Doing this with sufficient accuracy requires it to build an internal model of the world. We don't know exactly how human brains learn internally, but if we treat it as a black box, then like GPT4 it takes in words/texts and learns to output words/text that make correct inferences about the world.