PhD Fellowship and Postdoc in Explainable AI, LLM Security, and Mechanistic Interpretability University of Copenhagen in Denmark

University of Copenhagen is advertising PhD and postdoc opportunities in Explainable AI, LLM Security, Mechanistic Interpretability, and Natural Language Processing at the Department of Computer Science, Faculty of Science.

The PhD fellowship is titled “PhD fellowship in Mechanistic Interpretability for LLM Security”. The project is part of “A Mechanistic Framework for Mitigating the Susceptibility of LLMs to Learning False Information”, funded by Independent Research Foundation Denmark and led by Isabelle Augenstein and Pepa Atanasova. The research focus includes theoretical frameworks for LLM security, mechanistic interpretability methods, evaluation protocols, fact checking, cross-cultural learning, and responsible AI.

The postdoc position is titled “Postdoc in LLM Factuality Detection” and is also within the same project. The postdoc will work on characterising false information used by LLMs across the model lifecycle and developing mitigation methods against false information attacks.

Eligibility highlights: the PhD requires a relevant MSc degree or equivalent in Computer Science or a related field plus strong English skills; the ideal candidate has ML/NLP background, publications, or work experience. The postdoc requires a PhD degree or equivalent in Computer Science or a related field plus strong English skills; experience in NLP, LLM security, and/or explainable AI is preferred.

Funding and terms: both positions are full-time and offered for up to 3 years. The postdoc notice mentions salary supplement negotiation is possible. No stipend amount is stated.

Application deadline: 31 May 2026, 23:59 GMT+1. Applications must be submitted electronically via the University of Copenhagen job portal. Required materials include a cover letter, research statement or research plan, CV, diplomas/transcripts, publication list, and references.

Supervisors and contacts: principal supervisor Prof. Isabelle Augenstein and co-supervisor Pepa Atanasova, both at the University of Copenhagen, Department of Computer Science.