Itay Yona

-

Bio

Itay Yona is an AI security researcher and mechanistic interpretability specialist who founded MentaLeap and serves as its founder and principal investigator while also working as an AI security researcher at Google DeepMind.

Links

LinkedIn: LinkedIn
LessWrong: itay-yona

Organizations

MentaLeap

Founder & Principal Investigator

Projects

Cultivating a Mechanistic Interpretability Community

creator

Doublespeak: In-Context Representation Hijacking

Co-author

Friends and Grandmothers in Silico: Localizing Entity Cells in Language Models

Co-author

Soft Prompts are a Convex Set

Team member

Grants

No grants recorded.