Program Overview

MATS is an independent research and educational seminar program that connects talented scholars with top mentors in the fields of AI alignment, interpretability, and governance. The main goal of MATS is to help scholars develop as AI safety researchers.

Timeline

  • Applications Open

    Rolling

    Applications for future cohorts of MATS are open here. You can apply for the Summer 2024 and Winter 2024-25 programs now.

  • Applications Due

    March 24, 2024 [EDIT: Extended to April 7.]

    The final deadline for Summer 2024 applications is March 24.

    We suggest submitting your application by March 10 (the priority deadline) if possible since we will begin reviewing applications on March 11.

  • Application: Second Stage

    Late March

    Applicants who make it past the first stage will receive emails that allow them to indicate which mentors’ streams they wish to apply to. Applicants will receive the corresponding mentor selection problems, which are questions and tasks selected by mentors to evaluate your aptitude as a researcher. Note: some mentors request that you spend as long as ~10 hours on their selection problems.

  • Applicant Selection

    Late April

    Selected applicants will hear back about their acceptance into the program and anticipated mentor.

  • Research Phase

    June 17 - August 23

    During the 10-week research phase, scholars will collaborate with mentors to explore and work on a research project. During this time, scholars will write a short Scholar Research Plan and give a short presentation on their project work at the Scholar Symposium.

  • Extension Phase

    September 9 - December 20

    If scholars demonstrate promise as an independent researcher, MATS will consider them for the extension program which occurs for approximately four months immediately after the research phase. This occurs in both Berkeley, California and at the London Institute of Safe AI (LISA).

Program Phases

  • During the Research phase, each scholar spends ~1-2 hours/week working with their mentor, with more frequent communication via Slack. Scholars' research directions will typically be chosen through a collaborative process with their mentors, and scholars are expected to develop their independent research direction as the program continues. Approximately one month into the program, scholars are expected to write a short paper called the Scholar Research Plan outlining their projects’ threat model, theory of change, and project deliverables. At the end of the program scholars will give a brief presentation at the Scholar Symposium on project work conducted over the course of MATS.

    Educational seminars and workshops will be held 2-3 times per week. Previously, speakers have included Buck Shlegeris from Redwood Research, Adam Gleave from FAR AI, Neel Nanda from Google DeepMind, William Saunders from OpenAI, Andrew Critch from CHAI, Lennart Heim from GovAI, Ajeya Cotra from Open Philanthropy, and more.

    The extent of mentor support will vary depending on the project and the mentor. Scholars will also receive support from MATS’s Research Management team, who help to scope out and structure research direction.

  • Possible ongoing 4-month extensions for select scholars, potentially in Berkeley, California or London, UK. Pending mentor review, scholars who have received funding from the Long Term Future Fund will continue into the Extension phase to continue their research.

  • MATS aims to accelerate researchers who will:

    MATS alumni have gone on to publish safety research, join alignment organizations, including Anthropic and MIRI, and found an alignment research lab. You can read more about MATS alumni here.

Community at MATS

In contrast to doing independent research remotely, the MATS Research phase provides scholars with a community of peers. During the Research phase, scholars work out of a shared office, have shared housing, and are supported by a full-time Community Manager.

Working in a community of independent researchers gives scholars easy access to future collaborators, a deeper understanding of other alignment agendas, and a social network in the alignment community.

Previous MATS cohorts included regular lightning talks, scholar-led study groups on mechanistic interpretability and linear algebra, and hackathons. Other impromptu office events included group-jailbreaking Bing chat and exchanging hundreds of anonymous compliment notes. Scholars organized social activities outside of work, including road trips to Yosemite, visits to San Francisco, and joining ACX meetups.