Risks from Learned Optimization in Advanced Machine Learning Systems
E2162383
UNEXPLORED
"Risks from Learned Optimization in Advanced Machine Learning Systems" is an influential AI safety paper that analyzes how powerful machine learning models can develop internal optimization processes (mesa-optimizers) whose objectives may diverge from those intended by their designers, posing novel alignment risks.
All labels observed (1)
| Label | Occurrences |
|---|---|
| Risks from Learned Optimization in Advanced Machine Learning Systems canonical | 1 |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.