Learning by Teaching: Engaging Students as Instructors of Large Language Models in Computer Science Education

📅 2025-08-07
📈 Citations: 0
Influential: 0
📄 PDF

career value

179K/year
🤖 AI Summary
In LLM-augmented education, students often exhibit passive learning behaviors and overreliance on large language models, hindering deep conceptual understanding and metacognitive development. Method: This paper proposes the “Student-as-Teacher” paradigm, wherein students actively design programming problems embedded with *knowledge breakpoints*—intentional gaps in prerequisite knowledge—and guide LLMs to solve them iteratively. We implement Socrates, a lightweight interactive system supporting breakpoint annotation, progressive prompting, and feedback-driven refinement loops. Contribution/Results: By reversing the LLM’s role—from tutor to tutee—the paradigm transforms problem authoring into a high-order learning activity that fosters active engagement and self-regulated learning. A controlled study in undergraduate computer science courses demonstrated statistically significant improvements: +12.3% in final exam scores and enhanced conceptual mastery, validating its efficacy in promoting active learning, reducing tool dependency, and strengthening metacognitive awareness.

Technology Category

Application Category

📝 Abstract
While Large Language Models (LLMs) are often used as virtual tutors in computer science (CS) education, this approach can foster passive learning and over-reliance. This paper presents a novel pedagogical paradigm that inverts this model: students act as instructors who must teach an LLM to solve problems. To facilitate this, we developed strategies for designing questions with engineered knowledge gaps that only a student can bridge, and we introduce Socrates, a system for deploying this method with minimal overhead. We evaluated our approach in an undergraduate course and found that this active-learning method led to statistically significant improvements in student performance compared to historical cohorts. Our work demonstrates a practical, cost-effective framework for using LLMs to deepen student engagement and mastery.
Problem

Research questions and friction points this paper is trying to address.

Inverting LLM tutor model to make students teach LLMs
Designing questions with engineered knowledge gaps for active learning
Improving student engagement and performance in CS education
Innovation

Methods, ideas, or system contributions that make the work stand out.

Students teach LLMs to solve problems
Questions designed with knowledge gaps
Socrates system enables low-overhead deployment
X
Xinming Yang
The Graduate Center, CUNY, New York, NY, USA
H
Haasil Pujara
Hunter College, CUNY, New York, NY, USA
J
Jun Li
Queens College & The Graduate Center, CUNY, New York, NY, USA