GPT (Generative Pre-trained Transformer)

Probability Density Modeling

Fundamental objective of language models that learn to estimate the conditional probability P(token_t | tokens_<t) for each position in a sequence.

← Back