Research Scientist, Mechanistic Interpretability
Google DeepMindLondon, UK
AI safety & policy
Full-time
AI safety & policy
Full-time
Routes to impact
Direct high impact on an important cause
Description
Lead research on mechanistic interpretability in large language models, focusing on self-conceptions, theory of mind, and internal mechanisms shaping model outputs.
- Define and lead a novel mechanistic interpretability research agenda for LLMs
- Adapt and extend interpretability techniques for Gemini and other models
- Design experiments to identify internal vectors and mechanisms influencing outputs
- Collaborate with technical alignment teams on AI safety and capabilities initiatives
- Publish findings in papers, internal reports, and cross-disciplinary presentations
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.
