Blog
Join our mailing list to stay up to date on our work.
AI speeds up 3rd party safety researchers by ~1.6x (via self-reports)
Alignment
Research
Angel Martinez, Dewi Gould, Nicole Xing, Alejandro Aristizabal
Labs’ current alignment plans seem terrible
Alignment
Strategy
Daniel Tan, Andrew Draganov, Nicole Xing
Lessons from building an automated research scaffold
Alignment
Research
Alejandro Aristizabal, Josh Hills, Dewi Gould, Angel Martinez, Falko Galperin, Denis Federico Lim, Aleksandr Bowkis
Alignment Midtraining Cracks Under Pressure
Alignment
Research
Jonathan Bostock, Sid Baines, Daniel Tan, Andrew Draganov, Angel Martinez, David Africa
Obstacles to the scalable oversight of auto-alignment research
Alignment
Research
Sam Martin, Dewi Gould, Cameron Holmes, Jacob Pfau
Automated alignment runs are hard to study!
Alignment
Research
Alejandro Aristizabal, Andrew Draganov, Aleksandr Bowkis, Cameron Holmes
Debate with Self-Play Best-of-N Optimization
Alignment
Research
Dewi Gould, Sam Martin, Alejandro Aristizabal, Simon Marshall, Jacob Pfau
Your Model Organisms Might Be Fried
Alignment
Research
Daniel Tan, Jonathan Bostock, Andrew Draganov, Angel Martinez, Sid Baines, David Africa
Learnings from starting an AI safety research team
Organisation
Andrew Draganov, Erin Robertson