PaLM: Scaling Language Modeling with Pathways. (arXiv:2204.02311v4 [cs.CL] UPDATED)

Large language models have been shown to achieve remarkable performance
across a variety of natural language tasks using few-shot learning, which
drastically reduces the number of task-specific training examples needed to
adapt the model to a particular application. To further our understanding of
the impact of scale on few-shot learning, we trained a 540-billion parameter,
densely activated, Transformer language model, which we call Pathways Language
Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML …

arxiv language modeling palm scaling