This work was done during one weekend by research workshop participants and does not represent the work of Apart Research.
ApartSprints
Safety Benchmarks
Accepted at the 
 research sprint on 
Accepted at the 
Safety Benchmarks
 research sprint on 
July 3, 2023

AI & Cyberdefense

[unfinished] While hosting the hackathon, I had a few hours to explore safety benchmarks in relation to cyberdefence and mechanistic interpretability. I present a few project idea and research paths that might be interesting in the intersection between existential AI safety and cyber security.

By 
Esben Kran
🏆 
4th place
3rd place
2nd place
1st place
 by peer review