Agnieszka Karlińska

Head of the Large Language Models Department at the Center for AI Safety Research at NASK PIB, and project lead of HIVE AI – an initiative focused on the development and pilot implementation of Polish language models in public administration. Her work involves both building safe and trustworthy large language models and developing tools to detect harmful content, such as hate speech. She focuses particularly on evaluating the tendency of LLMs to generate toxic and discriminatory content, as well as mitigating biases at various stages of model development. She also led the PLLuM project on behalf of NASK PIB.