Stanford CRFM

The Center for Research on Fundamental Models (CRFM) at the Stanford Center for Artificial Intelligence (HAI) is an interdisciplinary initiative. This center aims to make fundamental advances in the research, development, and deployment of foundational models.

Basic models, such as GPT-3, BERT, and CLIP, are trained on large-scale data and can be adapted to a wide range of downstream tasks. These models are changing how AI systems are built and have significant social impacts.

CRFM’s mission includes the following aspects:

Research: Conduct interdisciplinary research to make underlying models more efficient, robust, interpretable, multimodal, and ethically compliant.

Artifacts: train and release basic models, codes, and tools, and ensure that the entire training process is replicable and scientifically rigorous.

Community: Invite universities, companies, and non-profit organizations to work together to develop professional practices for responsible training and deployment of foundational models.

The center brings together more than 175 researchers from more than 10 Stanford departments. The goal is to develop open, easy-to-use tools, as well as rigorous principles, for training and evaluating underlying models. This allows a more diverse set of participants to effectively critique and improve them.