Join our research sprints and create MVPs for research projects within evals, governance and interpretability to be invited for the Apart Lab program.
Projects are selected in collaboration with our senior mentors. You will receive an invitation if a new cohort is starting and your project looks promising.
During a meeting with one of the mentors, we decide the deadlines and first steps towards realizing your project!
The first two weeks of research are part of a trial period to see if your team's fit to the project makes sense. At the end of this stage, you have an outline of the research question and methodology.
In this stage, you make the first draft ready for review by other Lab members. Here, we consider how to make the research have real-world impact with you.
After this step, the draft is sent to a senior mentor, e.g. potential co-authors within alignment or internal senior mentorship.
Given the feedback from the last step, update the paper and get it into a submission-ready format.
Submit and (possibly) get your paper accepted after peer review! If not, we update the research and submit it to other conferences.
Get unique chances to join new research and build upon your existing research. Share your paper in talks and blog posts.
The fellowship aims to foster innovation in AI safety, build the fellows’ research skills, and prepare fellows for impactful roles in AI safety in academia or industry.
The typical primary research outcome is a published peer-reviewed paper. Depending on the project’s nature, the primary outcome may also be research software or a product prototype. Secondary outcomes could additionally include conference presentations, blog posts or other kinds of outreach or publication.
The fellowship aims for efficiency and timeliness, but the exact duration will depend on capacity of the involved team members and difficulty of the research project. We aim to complete the project (typically, with a publication-ready draft) within three to six months.
Support includes mentorship from experienced researchers, research project management, access to resources like GPUs, and a collaborative research environment with a cohort of like-minded Apart Lab fellows.
The mentorship includes research guidance from experienced AI safety researchers, as well as support via research project management. Mentorship and project management happen via regular meetings with the whole project team, as well as in a 1-1 setting on demand.
The mentors at Apart Lab are experienced researchers and professionals in the field of AI safety and related areas. Read more about our mentors here. Apart Lab also maintains connections with external researchers to provide additional feedback to our Apart Lab fellows.
Yes, the fellowship is designed to be completed remotely and independently of time zones, offering flexibility and accessibility to participants globally.
The fellowship primarily operates remotely, but there may be opportunities for in-person collaboration depending on the project and location. Please approach us individually about this so we can discuss the options.
The weekly time commitment is flexible but must allow meaningful progress on the research project, meaning typically at least 10 hours per week.
The research projects will be owned by Apart Lab fellows who will become the lead authors on resulting publications. As usual in academia, first authorship and coauthorship will be shared with anyone substantially involved in the research process depending on their contributions.
Yes, Apart Lab supports collaboration and co-authoring with external researchers, subject to the relevance and alignment with the research project.
We are currently applying for funding to support fellows with stipends bound to project deliverables; however, we do not have funding confirmed for this at the moment so cannot make any promises. As part of our mentorship, we will offer support for fellows to apply to outside funding sources wherever possible.
Participants will receive recognition for their involvement primarily via authorship on any publications. We also provide graduates of the Apart Lab fellowship with certificates of completion.
Research at Apart Lab focuses on core topics within AI safety, including model evaluations, model interpretability, multi-agent systems, and AI governance.
Communication and collaboration at Apart Lab happen asynchronously via Discord and other remote collaboration tools, as well as via regular (online) meetings within the project team, within the whole lab cohort and in 1-1 settings with mentors and team members.
Apart Lab helps setting up teams for efficient collaboration, with clear roles and responsibilities, owned and driven by the lab fellows themselves but with support and guidance from Apart Lab mentors.
Progress is monitored through written updates and regular check-ins, project milestones, and feedback sessions.
Feedback at Apart Lab happens in meetings with Apart Lab mentors as well as peer-to-peer within the project teams and the lab cohort.
There is no application. To join the fellowship, join one of our Apart Sprints. We invite the most promising teams to join the fellowship.
The fellowship is open to individuals with a strong interest in AI safety research and the required skills, as demonstrated during one of our Apart Sprints. No formal credentials are required.
Apart Lab’s explicit goal is to allow everyone across the globe to participate in AI safety research according to their capabilities. If you feel uneasy about joining Apart Lab because you feel you might not fit in, please get in touch at firstname.lastname@example.org so we can figure out how to best support you.
Yes, the fellowship provides networking opportunities through team collaboration on research projects, mentoring with senior researchers, and events and interactions with the wider Apart Lab community.