Tag
This paper introduces the Developmental Sentence Completion Test (DSCT) to evaluate Large Language Models' ability to recognize developmental cognitive stages in text, finding that models perform better on synthetic personas than on real human responses.