Description
Title: Low-Resource Preference Adaptation for LLMs
Abstract:
Adapting large language models to user-specific preferences is often constrained by the cost of
human annotation, making preference optimisation impractical in low-resource settings where
preferences cannot be reliably labelled by LLMs
themselves, e.g., due to cultural, subjective, or personalised contexts. I will present ways for investigating how language models encode preference information in their intermediate representations, finding that activations from chosen and rejected responses form distinct clusters across layers, even in pretrained models.
Bio:
Meriem Beloucif is an Assistant Professor in Computational Linguistics at Uppsala University. Her research focuses on large language models for low-resource languages, including neural machine translation, lexical semantics, multilingual resource development, and the systematic evaluation of language models. Before, she worked on comparative QA and Information Retrieval at Hamburg University and spent a year at Copenhagen University as a Postdoc. She has a PhD in Computer Science from Hong Kong University of Science and Technology.