Generating Educational Materials with Different Levels of Readability using LLMs
This addresses the need for adaptable educational content generation, though it is incremental as it builds on existing LLM capabilities.
The study tackled the problem of rewriting educational materials to specific readability levels using large language models, finding that few-shot prompting improves performance with LLaMA-2 70B better at achieving difficulty ranges and GPT-3.5 better at preserving meaning, but manual inspection revealed issues like misinformation.
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.