WhiteBox Research will have at least four fellows spending ≥10 h/wk on an alignment-oriented project by EoY 2024.
Basic
3
989
Dec 31
79%
chance

WhiteBox Research is currently holding an interpretability fellowship in Manila. 11 participants are currently in our fellowship’s guided research phase, 6 of whom will use Neel Nanda's 200 Concrete Open Problems to rapidly upskill in mechanistic interpretability. We are currently funded by the Long-Term Future Fund and Manifund.

Some of the fellows in our first cohort include:
- an IOI silver medalist
- an Iranian Geometry Olympiad bronze medalist
- an AI engineering technical manager and former trainer for the Philippine IOI team

Other fellows have also won first and third places in recent Apart Hackathons.

Said project must last at least four weeks, and may not necessarily be funded by any external funding source or grantmaker.

Get Ṁ1,000 play money