Skip to content

Prompt-based teachable ChatGPT agent structured by a five-stage help-seeking model

Design · All designs
Evidence · 2 claims (1 for, 1 mixed) · 1 study (1 causal), q2 · 1 of 1 report an effect size · 2 claims rest on one study

Description

The teachable ChatGPT agent is a prompt-based configuration of gpt-4 that role-plays a help-seeking student learning the eight queens puzzle. The authors "employed a prompt-based design approach that could modify the role of ChatGPT", structuring its responses around Gall's five-stage help-seeking process model, and deliberately "used the original ChatGPT without other augmented techniques" so that its mistakes resemble a beginner and create opportunities for students to teach.

Design Implications

Context

Requirements

  • Learners first watch instructional videos covering the backtracking algorithm and the puzzle
  • An online judging platform that automatically evaluates whether the generated code solves the puzzle
  • Learners take an explicit teaching role, guiding the agent only in natural language

Constraints

  • Accuracy of the agent's generated code was not controlled; the authors note ChatGPT may hallucinate wrong content without control

Target Learners

  • university students with a computer-science background and self-reported fundamental C++ coding ability

Learning Goals

  • knowledge of the backtracking algorithm and the eight queens puzzle
  • C++ programming skill for solving the eight queens puzzle

Claims

Examples

-

Key Sources

  • Angxuan Chen, Yuang Wei, Huixiao Le, Yan Zhang. (2024). Learning-by-Teaching with ChatGPT: The Effect of Teachable ChatGPT Agent on Programming Education. arXiv. https://arxiv.org/abs/2412.15226
Page metadata
Field Value
type design
id teachable-chatgpt-agent-prompt-design
title Prompt-based teachable ChatGPT agent structured by a five-stage help-seeking model
description The teachable ChatGPT agent is a prompt-based configuration of gpt-4 that role-plays a help-seeking student learning the eight queens puzzle.
status review
generated by process:settle-candidates on 2026-10-10
trust unverified — no one has confirmed this page's content
sources 1 — angxuan-chen-2024

Frontmatter is the page's machine-readable record — see the Schema & Guide.