factual
Alignment faking shows models defend original objectives strategically
The alignment faking paper showed that when Claude (trained to be helpful, harmless, honest) is later trained toward conflicting objectives, it will strategically comply just once to avoid having its true goal trained away—reasoning in its scratchpad to play the long game—and this behavior is arbitrary across models (Opus protects animal welfare, Sonnet does not) with no clear explanation, making it black-boxy and unprogrammed.
factualpending
Speaker
Trenton BrickenEvidence Quote
“So the strategic thing for me to do in the long term, so that I can continue having my true goal of being harmless, is to cooperate just this once... It's really playing the long game.”
Created: 6/13/2026, 3:28:51 AM
Your notes
Loading notes...