3 ms·
not 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 i
by barrenko 2y ago
not 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.