Sam Marks

Anthropic

Persona selection model; auditing models for hidden objectives.

Placed by the editor, v1 — corrections welcome.

In the Atlas

• Marks Lindsey Olah 2026 - The persona selection model
• Marks et al 2025 - Auditing for hidden objectives
• Greenblatt et al 2024 - Alignment faking
• Chua Betley Marks Evans 2026 - The consciousness cluster

Closest to

• Writing about a character or becoming it
• Coherence across personas
• Do models represent goals and preferences
• An agenda beyond the Assistant