Research Associate, Unverbalized Dangerous Capabilities in Frontier Models
Imperial College LondonLondon, UK
AI safety & policy
Full-time
AI safety & policy
Full-time
Deadline
2026-11-22
Salary
£50,733 to £59,484 per annum
Routes to impact
Direct high impact on an important cause
Description
Design and test methods to detect hidden dangerous capabilities in frontier models within a postdoctoral safety and alignment research role.
- Develop algorithms to detect and defend against unverbalized dangerous capabilities
- Conduct independent and collaborative research on frontier models and Chain-of-Thought monitoring
- Publish results in venues such as NeurIPS, ICML, ICLR, ACL, and EMNLP
- Implement research ideas and contribute to lab projects, meetings, and presentations
- Fixed term for 12 months to start ASAP and before 8 th April 2027
This text was generated by AI. If you notice any inconsistencies, please let us know using this form.
