"Pre-trained" is a little bit historical. Initially, the ability for the model to continue text was thought of as just a prerequisite for a more specialized task: inference (finding logical connections between phrases), classification (for instance, guessing the number of stars in a hotel ratin...
Initially, the ability for the model to continue text was thought of as just a prerequisite for a more specialized task: inference (finding logical connections between phrases), classification (for instance, guessing the number of stars in a hotel rating from the text of the review), machine ...