Daan R. Henselmans

LLMs lack moral competence for alignment, new papers find

Two new research papers explore the limitations of current Large Language Models (LLMs) in handling moral reasoning and alignment. The first paper argues that LLM agents lack the fundamental "moral competence" required …