Leading Mechanistic Interpretability Research

Acceler8 Talent; San Francisco (CA); On-site; USD 190,000 - 320,000; Full time

Job summary

Acceler8 Talent is seeking a very senior researcher in San Francisco or Seattle to set research direction and develop original interpretability methods for large language models. You would operate with significant independence and have the freedom to shape research directions while mentoring others around you.

The focus is on mechanistic interpretability: understanding internal representations, tracing and manipulating circuits, and uncovering latent concepts to make frontier models more

Responsibilities

• Set research direction for mechanistic interpretability of large language models.
• Develop original interpretability methods and circuits.
• Mentor team members and collaborate with researchers.

Skills

Mechanistic interpretability

Research leadership

Independence

Job description

I’m working with an early-stage AI research lab focused on mechanistic interpretability and alignment for large language models.

They’re hiring a very senior researcher in San Francisco or Seattle who can set research direction, develop original interpretability methods and operate with significant independence. You'd be given the freedom to shape research direction and mentor others around you.

The focus is firmly on mechanistic interpretability: understanding what models represent internally, tracing and manipulating those representations, uncovering circuits and latent concepts, and developing new ways to make frontier models more understandable and steerable.

There’s no requirement to be a people manager.