Temperature is a common LLM hyperparameter that controls the randomness of the model's output. This post explains how temperature sampling works.
should "randomness is reduced and the model is more likely to pick a high-probability word" be "randomness is reduced and the model is more likely to pick a low-probability word"?
should "randomness is reduced and the model is more likely to pick a high-probability word" be "randomness is reduced and the model is more likely to pick a low-probability word"?