Marius Hobbhahn

Apollo Research

Marius took part in MATS Winter 2022/23 Cohort under the mentorship of Evan Hubinger (Anthropic). He published multiple pieces on mechanistic interpretability on LessWrong including work on maximum data dimension and double descent. He is currently the CEO and Director of Apollo Research, a new London-based technical alignment organization. Previously, he did a Ph.D. in Machine Learning and conducted independent alignment research. Read more on his website.

The Winter 2022-23 cohort supported 58 scholars with 17 mentors including researchers from Anthropic, MIRI, ARC, Redwood Research, and other leading organizations. This cohort introduced the Scholar Support team to provide research coaching and unblocking assistance to scholars throughout the program. The program ran 6 weeks online followed by 2 months in-person in Berkeley and featured scholar-led activities including study groups on mechanistic interpretability and linear algebra, weekly lightning talks, and workshops on research tools and technical writing.Notable alumni from this cohort include Marius Hobbhahn, who founded Apollo Research and published work on mechanistic interpretability; Asa, who co-authored papers on measuring situational awareness and the "reversal curse" in large language models; and Jesse Hoogland, who founded Timaeus and developed the developmental interpretability research agenda.

Apollo almost certainly would not have happened without MATS. One of the core reasons why starting an organization is hard is because the founding members need to know and trust each other. It is often hard to find people with similar agendas that you also personally enjoy working with in a systematic manner. MATS implicitly created such an environment because it enabled many of us to understand what everyone else is working on, get to know them personally and see their research progress without having to commit to anything in particular.

Marius Hobbhahn