Skip to main content Skip to navigation

RECIPE4U: Student-ChatGPT Interaction Dataset in EFL Writing Education

Project Overview

The document introduces RECIPE4U, a comprehensive dataset developed from a semester-long study of 212 college students enrolled in English as a Foreign Language (EFL) writing courses, which explores the application of generative AI, specifically ChatGPT, to improve writing skills through essay revisions. This dataset encompasses conversation logs, students' intents, satisfaction ratings, and the history of essay edits, providing valuable insights into how large language models (LLMs) are utilized in educational settings. The findings highlight the positive impact of AI-assisted writing support on student engagement and skill enhancement, while also underscoring the necessity for systematic research to better understand LLM integration in education. The document emphasizes the potential of generative AI to transform EFL instruction and improve learning outcomes, advocating for further exploration of AI technologies in educational contexts to optimize their effectiveness and address challenges in language learning.

Key Applications

RECIPE4U dataset for EFL writing education

Context: Higher education, specifically for EFL learners in writing courses.

Implementation: A semester-long experiment with students engaging in dialogues with ChatGPT to revise essays.

Outcomes: Improvement in essay quality, understanding of student interaction patterns, and establishment of baseline models for intent detection and satisfaction estimation.

Challenges: Most LLM tools are not designed for education; understanding long-term usage patterns remains a challenge.

Implementation Barriers

Technical Barrier

Most LLM-based tools are not originally designed for educational purposes.

Proposed Solutions: Explore potential and actual utilization patterns of LLMs within educational settings.

Research Gap

Lack of comprehensive, long-term analyses of real-world usage of LLMs in education.

Proposed Solutions: Conduct systematic studies to capture semester-long interactions between students and AI.

Project Team

Jieun Han

Researcher

Haneul Yoo

Researcher

Junho Myung

Researcher

Minsun Kim

Researcher

Tak Yeon Lee

Researcher

So-Yeon Ahn

Researcher

Alice Oh

Researcher

Contact Information

For information about the paper, please contact the authors.

Authors: Jieun Han, Haneul Yoo, Junho Myung, Minsun Kim, Tak Yeon Lee, So-Yeon Ahn, Alice Oh

Source Publication: View Original PaperLink opens in a new window

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai

Let us know you agree to cookies