3 ms·
I still don't get the reinforcement part here. Wouldn't that be normal training against the data set? Like how would you modify the normal MNIST training to be
by singularity2001 2y ago
I still don't get the reinforcement part here. Wouldn't that be normal training against the data set? Like how would you modify the normal MNIST training to be reinforcement learning
- barrenko 2y agonot an expert - yes, what would usually just be called training, with LLMs here is called RL. You do end up writing a sort of a reward function, so I guess it is RL.
- hodapp 2y agoYou are right; the advanced in DeepSeek-R1 used RL almost solely because of the chain-of-thought sequences they were generating and training it on.