Call for Papers

Language Models are few short learners is an important paper in the space of GenerativeAI and Natural Language Processing. It introduced GPT-3 and showed the capability of large language models to generalize as task-agnostic learners.

The paper sowed the seeds for building NLP applications by prompting large language models with zero-shot, one-shot, and few-shot learning prompts. This was a huge advancement from task-specific modeling and also closer to how the human brain works by applying past learning to new data.

GPT-3 used the similar but scaled-up(100x) model architecture as GPT-2 except for the use of Sparse Attention (introduced in the Sparse Transformer paper).

The paper talks in great detail about the result and impact.

In this session, I will provide a condensed and simplified understanding of the key points and takeaways from this long paper.

Speaker Intro
Simrat has a career spanning over a decade in the AI ML space, specializing in Natural Language Processing.
Currently spearheading AI product strategy at Hasura and has led AI teams at renowned organizations such as VMware, FI Money, and Nirvana Insurance in the past.