Hi, I'm Chloe.

I'm currently a Research Fellow at Anthropic. I'm concerned that advanced AI systems may develop the ability to secretly pursue harmful goals of their own. I've been fortunate to collaborate with Joe Benton, Jon Kutasov, and Tyler Tracy on AI control research.

Previously:

• • Trained a video generation transformer.
• • First author on DafnyBench with Nada Amin and Max Tegmark.
• • Co-authored VerMCTS, a verifier-guided reasoning and synthesis approach. Pitched related startup idea to Sam Altman, later featured in Harvard Business School case study.
• • 2018 AI Grant Fellow (Nat Friedman & Daniel Gross), where I got my start doing independent AI research.