GPT-2 and GPT-3
Miles discusses the evolution from GPT-2 to GPT-3, highlighting the shift in focus from responsible publication to deployment via an API. He explains how both models are transformer-based and capable of performing a variety of tasks by learning from diverse internet text. The introduction of few-shot learning in GPT-3 allows the model to adapt to new tasks with minimal examples, sparking significant advancements in language model research and applications.In this clip
From this podcast
Super Data Science: ML & AI Podcast with Jon Krohn
SDS 597: A.I. Policy at OpenAI — with Miles Brundage
Related Questions