3 Mechanisms Underlying Emergent Abilities in Generative Models
October 22, 2024
Summary
Hidenori Tanaka (Harvard’s Physics of Intelligence program) presented "3 Mechanisms Underlying Emergent Abilities in Generative Models" at FAR Labs, as part of the FAR Seminar series. His research crosses the intersection of physics, neuroscience, and psychology to understand AI creativity and ensure AI safety.
Key Highlights:
- How emergent AI abilities evolve
- The role of foundational science in AI safety
SESSION Transcript
