Large language models (LLMs) offer an opportunity to make large-scale changes to educational content that would otherwise be too costly to implement. The work here highlights how LLMs (in particular GPT-4) can be prompted to revise educational math content ready for large scale deployment in real-world learning environments. We tested the ability…

This paper describes and reports on the performance of six related artificial neural networks that have been developed for the purpose of readability analysis. Two networks employ counts of linguistic variables that simulate a traditional regression-based approach to readability. The remaining networks determine readability from "visual…