AI Safety and Alignment

Channel: #safety-and-alignment

Co-leads: 

• Alif - @biggmon on Discord

• Sunitha - @prisca6117 on Discord

Goal: 

• Each session covers different topics in safety and alignment. The goal is for you to come out of each session with at least a high-level understanding of what we discussed. 

• Topics include (but are not limited to) RLHF, inner & outer misalignment, goal mis-generalization, inverse RL, scalable oversight, interpretability. 

Logistics:

• Everyone is welcome to join! A basic ML background is sufficient. 

• We meet on Thursday at 10 am PST (on a bi-weekly basis). 

Recent Recordings

May 30, 2024