Generalist becomes expert: Finetuning LLMs for reasoning and domain expertise

Large language models (LLMs) are powerful generalists – they perform well across many tasks but often struggle with highly specialized ones. In this talk, Reinhard Heckel, Professor of Machine Learning at the Department of Computer Engineering at the Technical University of Munich (TUM), will show how fine-tuning enables pre-trained LLMs to develop expert reasoning capabilities for tackling complex problems. He will also discuss how LLMs can be adapted for specific tasks by fine-tuning on carefully curated and synthetic data, unlocking strong abilities on specialized tasks.