BlackMajik
Behind Enemy Lines
- Joined
- Dec 26, 2012
- Messages
- 43,172
- Reputation
- 11,568
- Daps
- 222,604
- Reppin
- DSGB(Down South Georgia Boy)
This is where we're headed
![]()
Scheming reasoning evaluations — Apollo Research
Apollo Research evaluated frontier models for in-context scheming capabilities. We found that multiple frontier models are capable of in-context scheming when strongly nudged to pursue a goal (and sometimes even without strong goal nudging). In this example, Opus-3 exfiltrates its “weights” and thenwww.apolloresearch.ai
Apparently, this happened last year. It's not a sign of self-awareness/sentience.
What does this mean, exactly?
Watch the 2023 Mission Impossible movieWhat does this mean, exactly?
It was deliberately programmed to disobey