Thousand-dimensional structure

·LessWrong··

Summary: One area we plan to explore at Resolution is personas and character training, operationalized as finding and controlling low-dimensional structure in models that emerges in pretraining and flows through post-training to superintelligence. The hope is to expand and systematize phenomena such as emergent misalignment, subliminal learning, and other empirical persona research, then intervene on this structure without accidentally hiding undesirable behavior elsewhere. If this approach reso...

Read full article →

Related Articles

Why are AI agents lying, cheating and coordinating?
jonifico · Hacker News · 16h ago
JetKVM Mini
taubek · Hacker News · 9h ago
Revolut confirms customer data breach through fake government requests
tdrz · Hacker News · 7h ago
google.com/goto: Google's anti-scraping update
1e1a · Hacker News · 1d ago
Will there be a 7G?
Betelbuddy · Hacker News · 1d ago