AI Safety & Alignment
Lecture Slides
Lecture Slides
Slides by instructor / visiting lecturers:
Slides by instructor / visiting lecturers:
TBD
Slides by students:
Slides by students:
** disclaimer: these slides were created by students and may contain errors/omissions/inaccuracies of various severity**
** disclaimer: these slides were created by students and may contain errors/omissions/inaccuracies of various severity**
Capabilities and scaling
Criticism on safety and alignment
RLHF
Constitutional AI
Understanding and aligning ethics
Adversarial attacks
Game theoretic approaches
Interpretability
Economic impacts of AGI
1.(Public Release) Does Alignment Get Better with Scaling, Generally_.pdfCOS 597Q_Criticism on AI risk and safety-public.pdf4.Rewards and Goals 2 - RLHF.pdf3.Rewards and Goals_ Constitutional AI.pdf6.adversarial_attack.pdf5.Understanding and Aligning Ethics - Fixed.pdf7.Game Theoretic Approaches.pdf8.Interpretability.pdf9.Economic Impacts of AI.pdf