Agent Foundations
Scalable OversightactiveTheoretical foundations for reasoning about goal-directed AI systems (MIRI-style research).
First Proposed: 2014 (MIRI)
Cluster: Scalable Oversight
Tags
function:specificationscope:field
Sub-Areas1
| Name | Status | Orgs | Papers |
|---|---|---|---|
| Natural AbstractionsHypothesis that natural abstractions generalize across observers, providing a basis for alignment. | active | 0 | 0 |