A Framework for Fine-Grained Complexity Control in Health Answer Generation (Ferreira et al., ACL 2025)
RQ1: What combination of metrics best captures the complexity of medical text?
RQ2: How do we create a good training dataset for complexity-controlled text generation?
RQ3: How can we design a language model that reliably generates health answers with adjustable complexity levels?