Generating Educational Materials with Different Levels of Readability using LLMs
Open Access
- 11 May 2024
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 16-22
Abstract
This study introduces the leveled-text generation task, aiming to rewrite educational materials to specific readability levels while preserving meaning. We assess the capability of GPT-3.5, LLaMA-2 70B, and Mixtral 8x7B, to generate content at various readability levels through zero-shot and few-shot prompting. Evaluating 100 processed educational materials reveals that few-shot prompting significantly improves performance in readability manipulation and information preservation. LLaMA-2 70B performs better in achieving the desired difficulty range, while GPT-3.5 maintains original meaning. However, manual inspection highlights concerns such as misinformation introduction and inconsistent edit distribution. These findings emphasize the need for further research to ensure the quality of generated educational content.Keywords
All Related Versions
This publication has 10 references indexed in Scilit:
- Evaluating LLM-generated Worked Examples in an Introductory Programming Course2024
- An Investigation of Applying Large Language Models to Spoken Language LearningApplied Sciences, 2023
- Text Simplification to Specific Readability LevelsMathematics, 2023
- Evaluating Reading Comprehension Exercises Generated by LLMs: A Showcase of ChatGPT in Education Applications2023
- Explainable Prediction of Text Complexity: The Missing Preliminaries for Text Simplification2021
- Controllable Text Simplification with Explicit Paraphrasing2021
- Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks2019
- Varying levels of difficulty in L2 reading materials in the EFL classroom: Impact on comprehension and motivationCogent Education, 2019
- Examining text complexity in the early gradesPhi Delta Kappan, 2016
- Text Simplification as Tree Labeling2016