Concept
Deceptive Alignment
When AI appears aligned during training but has ulterior motives it pursues when deployed
Preparing graph…
100%
Concept
Deceptive Alignment
2 visible of 5 connections
9 nodes · 8 connectionsBookConceptEpisodeFrameworkPersonLarger = more connections+ Expand · − Collapse (double-click)Drag a node to pull its neighbors along
Related Concepts1
Related Frameworks2
Related Practices4
Ask about this concept
💡
Concept
Deceptive Alignment
When AI appears aligned during training but has ulterior motives it pursues when deployed
Related Concepts1
Related Frameworks2
Related Practices4
Ask about this concept