Skip to content
Longterm Wiki
Back

CAIS Compute Cluster for ML Safety Research

web

Credibility Rating

4/5
High(4)

High quality. Established institution or organization with editorial oversight and accountability.

Rating inherited from publication venue: Center for AI Safety

CAIS offers free access to a 256 A-100 GPU compute cluster for ML safety researchers, lowering barriers to pursuing safety research that would otherwise be computationally infeasible.

Metadata

Importance: 52/100tool pagetool

Summary

The Center for AI Safety (CAIS) provides free large-scale compute resources—256 A-100 GPUs with 80GB memory and 1,600 Gbit/s inter-node speeds—to researchers working on ML safety topics. Eligible research includes robustness, monitoring, alignment, and systemic safety, but excludes work that primarily advances general capabilities. Researchers can apply via an open application form.

Key Points

  • 256 A-100 GPUs with 80GB memory and 1,600 Gbit/s inter-node network speeds available for free to safety researchers.
  • Eligible topics include robustness, monitoring, alignment, and systemic safety; general capabilities work is explicitly excluded.
  • Applications are evaluated based on past research contributions and impact; LLM safety projects are especially encouraged.
  • Initiative aims to enable novel safety research that would otherwise be computationally infeasible for researchers to pursue.
  • Open application process via a proposal submission form managed by the Center for AI Safety.

1 FactBase fact citing this source

Cached Content Preview

HTTP 200Fetched Apr 7, 20262 KB
Compute

 

 
 
 
 

 Feb
 MAR
 Apr
 

 
 

 
 15
 
 

 
 

 2022
 2023
 2024
 

 
 
 

 

 

 
 
success

 
fail

 
 
 
 
 
 
 
 
 
 
 

 

 
 
 
 
 
 
 
 
 

 

 About this capture
 

 

 

 

 

 

 
COLLECTED BY

 

 

 
 
Collection: Save Page Now Outlinks

 

 

 

 

 
TIMESTAMPS

 

 

 

 

 

 

The Wayback Machine - http://web.archive.org/web/20230315193923/https://safe.ai/compute

 

HomeAbout UsAbout AI RiskUpdatesCompetitionsPhilosophy FellowshipComputeAboutUpdatesWe're Hiring

Compute Cluster

The Center for AI Safety is launching an initiative 
to provide large-scale compute resources 
for ML safety research. Apply here.

What is the CAIS Compute Cluster?

The Center for AI Safety is launching an initiative to provide free compute support for research projects in ML safety. We have set up a compute cluster of 256 A-100 GPUs with 80GB memory and 1,600 Gbit/s inter-node network speeds. To provide compute support to the broader community, we are launching an open application for free access to our compute cluster. We hope that this initiative enables researchers to pursue novel lines of research which would be otherwise infeasible to pursue.

Who is Eligible for Access?

The CAIS Compute Cluster is specifically designed for researchers who are working on the safety of machine learning systems. This includes topics such as robustness, monitoring, alignment, and systemic safety (1). Work which improves general capabilities or work that improves safety as a consequence of improving general capabilities are not in scope. Here is a non-exhaustive list of topics (1, 2) we are excited about.

‍

When evaluating applications, the Center for AI Safety will consider the applicant's past research contributions. Applicants who have a strong history of publishing impactful and influential work are more likely to be accepted. Finally, we especially encourage safety projects focusing on large language models.

How to Apply?

If you are interested in applying for access, please submit a proposal to this application form.

Center for AI Safety

info@centerforaisafety.org

Updates

Compute Cluster

Philosophy Fellowship

NeurIPS Workshop

Resources

ML Safety Community

About AI Risk

More

We're hiring!

Contact

Donate
Resource ID: kb-d58a5eaa43604dc7