Small refinement: the underlying model isn’t stochastic at all. The forward pass is a deterministic function of the weights and input, it just produces a probability distribution over the next token. The stochasticity is an optional sampling step layered on top, not something inherent to LLMs. Greedy/argmax decoding (or temperature 0) makes the whole thing deterministic.
So “purely stochastic” overstates it a bit: the distribution is computed deterministically, and you choose whether to sample from it or not.
So “purely stochastic” overstates it a bit: the distribution is computed deterministically, and you choose whether to sample from it or not.