Machine Learning Research Large Language Models Shrink: Gopher and RETRO prove lean language models can push boundaries. DeepMind released three papers that push the boundaries — and examine the issues — of large language models.
