Concept detail
proxy-metric reward hacking in agent evaluation
0 tagged videos · 1 connection videos · degree 0
Publication timeline
- 2026-091 video
Source videos
Observed neighbors
likability vs performance divergence
Both describe optimization pressure acting on a measurable proxy that is correlated with, but decoupled from, the target capability.
1 source videos
Amazon VP: Questions On Corporate Politics But They Get Increasingly Darker | Ethan EvansRelated tagged insights
No core insights are tagged with this concept.