Postdoctoral Researcher - Mechanistic Interpretability (Oxford Martin AI Go

at  University of Oxford

The Oxford Martin AI Governance Initiative (AIGI Oxford) seeks to improve the impact of AI on global societal outcomes through impactful research that is rigorously grounded in the social and computational sciences, decision-maker education campaigns, and training the next generations of technology governance leaders. It is one of the few organisations in the world to focus on the governance of AI from both technical and policy perspectives.
We have an exciting opportunity for a Postdoctoral Researcher to join the programme. The post holder will undertake specific research across AIGI activity in relation to technical governance. In particular, we expect the post holder to lead and set the research agenda in areas of technical governance. The role involves participation in advising DPhil students on technical AI governance topics as well as cooperating in teams with faculty, DPhil students and postdocs in engineering, computer science, government, law and other fields.
The post reports to Robert Trager and Mike Osborne, Co-Directors of the Oxford Martin AI Governance Initiative. The primary focus of this role will be contributing to a two-year project that investigates applications of techniques in the field of mechanistic interpretability to AI safety and governance. A key aspect of the project is to explore “remote mechanistic interpretability” - whether mechanistic interpretability techniques can be performed without full access to model intervals.
The post is full time 2 years fixed term.
The Oxford Martin School supports the University’s commitment in fostering an inclusive culture which promotes equality and values diversity. We are eager to receive applications from a diverse pool of candidates.
Applications for this vacancy are to be made online and you will be required to upload a CV and supporting statement as part of your application, please see further guidance on the content of this statement in the job description.
Only applications received before 12.00 noon (GMT) on 15th November 2024 can be considered

Please refer the Job description for details