Postdoc in LLM Factuality Detection

About the role

We are looking for a postdoctoral researcher to join the CopeNLU group at the University of Copenhagen. The position focuses on developing a mechanistic framework to detect and mitigate false information in large language models (LLMs) as part of an Independent Research Foundation Denmark funded project.

Key responsibilities

• Analyse the spectrum of false information that LLMs encounter throughout their lifecycle.
• Design and implement novel theoretical frameworks for LLM security.
• Develop mechanistic interpretability methods and new evaluation protocols.
• Collaborate with project leads Isabelle Augenstein and Pepa Atanasova, as well as external partners such as NVIDIA.
• Contribute to teaching and supervision activities on a limited basis.

Required profile

• PhD in Computer Science or a closely related field.
• Strong written and oral English communication skills.
• Demonstrated research record through publications in relevant areas.
• Experience in Natural Language Processing, LLM security, or Explainable AI.

Required skills

• Natural Language Processing
• LLM Security
• Explainable AI

What we offer

• Opportunity to work in a vibrant, internationally recognised NLP research environment.
• Access to state‑of‑the‑art resources and collaborations with industry partners.
• Integration into the Pioneer Centre for AI and its interdisciplinary research community.