Ana Kavramlar
Top-P (Nucleus Sampling)
A sampling strategy that limits token selection to the smallest set covering a cumulative probability threshold.
Top-P sampling (also called nucleus sampling) considers only the tokens whose cumulative probability mass reaches the threshold P. For example, top-p=0.9 samples from whatever tokens together account for 90% of predicted probability. It is typically used alongside temperature and offers another axis for controlling output diversity.