Evaluating LLMs on Generating Age-Appropriate Child-Like Conversations

Syed Zohaib Hassan, Pål Halvorsen, Miriam S. Johnson +1

Large Language Models (LLMs), predominantly trained on adult conversational data, face significant challenges when generating authentic, child-like dialogue for specialized applica…