WritingOpenAIOpenAIpublished Feb 4, 2021seen 6d

Understanding the capabilities, limitations, and societal impact of large language models

Open original ↗

Captured source

source ↗

Understanding the capabilities, limitations, and societal impact of large language models | OpenAI

February 4, 2021

Publication

Understanding the capabilities, limitations, and societal impact of large language models

Read paper

Loading…

Share

Abstract

On October 14th, 2020, researchers from OpenAI, the Stanford Institute for Human-Centered Artificial Intelligence, and other universities convened to discuss open research questions surrounding GPT‑3, the largest publicly-disclosed dense language model at the time. The meeting took place under Chatham House Rules. Discussants came from a variety of research backgrounds including computer science, linguistics, philosophy, political science, communications, cyber policy, and more. Broadly, the discussion centered around two main questions: 1) What are the technical capabilities and limitations of large language models? 2) What are the societal effects of widespread use of large language models? Here, we provide a detailed summary of the discussion organized by the two themes above.

  • GPT
  • Ethics & Safety
  • Community & Collaboration
  • Language

Authors

Alex Tamkin, Miles Brundage, Jack Clark, Deep Ganguli

Related articles

View all

Democratic inputs to AI grant program: lessons learned and implementation plansSafetyJan 16, 2024

Building agricultural database for farmersB2B StoryJan 12, 2024

Creating websites in minutes with AI Website BuilderB2B StoryMay 29, 2025