Large Language Models, GPT-3: Language Models are Few-Shot Learners | Towards Data Science

Efficiently scaling GPT from large to titanic magnitudes within the meta-learning framework