[论文解读] From Code-Centric to Concept-Centric: Teaching NLP with LLM-Assisted "Vibe Coding"

Hend Al-Khalifa|arXiv (Cornell University)|Feb 2, 2026

Computational and Text Analysis Methods被引用 0

一句话总结

该论文提出了 Vibe Coding，一种在 NLP 实验室中允许使用 LLM 辅助编码的教学法，同时通过提示日志与反思来评估学生的概念理解，尽管在时间和核验方面存在挑战，但展现出积极的参与度与感知的公平性。

ABSTRACT

The rapid advancement of Large Language Models (LLMs) presents both challenges and opportunities for Natural Language Processing (NLP) education. This paper introduces ``Vibe Coding,'' a pedagogical approach that leverages LLMs as coding assistants while maintaining focus on conceptual understanding and critical thinking. We describe the implementation of this approach in a senior-level undergraduate NLP course, where students completed seven labs using LLMs for code generation while being assessed primarily on conceptual understanding through critical reflection questions. Analysis of end-of-course feedback from 19 students reveals high satisfaction (mean scores 4.4-4.6/5.0) across engagement, conceptual learning, and assessment fairness. Students particularly valued the reduced cognitive load from debugging, enabling deeper focus on NLP concepts. However, challenges emerged around time constraints, LLM output verification, and the need for clearer task specifications. Our findings suggest that when properly structured with mandatory prompt logging and reflection-based assessment, LLM-assisted learning can shift focus from syntactic fluency to conceptual mastery, preparing students for an AI-augmented professional landscape.

研究动机与目标

• Motivate NLP education to balance AI-assisted coding with deep conceptual understanding.
• Introduce Vibe Coding as a three-component framework (sanctioned LLM use, mandatory prompt logging, reflection-based assessment).
• Evaluate the approach in a senior undergraduate NLP course with 19 students.
• Assess student engagement, perceived assessment fairness, and challenges of LLM-assisted labs.
• Investigate transfer of Vibe Coding skills to final project work.

提出的方法

• Implement Vibe Coding in a 12-week senior NLP course with 7 labs and a four-phase final project.
• Allow and encourage LLMs for code generation during labs; require prompt logs and reflection-based assessments.
• Assess via a 5-point Likert questionnaire covering course experience, Vibe Coding process, assessment structure, and project work.
• Analyze quantitative means and standard deviations and conduct thematic analysis on open-ended feedback.
• Labs cover NLP topics including tokenization, POS/NER, text classification, n-grams, embeddings, transformers fine-tuning, and in-context learning.

实验结果

研究问题

• RQ1 Structured 的 LLM 辅助编码（Vibe Coding）如何影响 NLP 实验室中的学生参与度与概念学习？
• RQ2 当评估优先考虑概念反思而非代码质量时，学生对评估公平性的感知为何？
• RQ3 当学生依赖 LLM 进行实验实现时，出现了哪些挑战？
• RQ4 学生如何将 LLM 辅助编码技能迁移到独立的项目工作？

主要发现

• 学生报告高参与度且认为课程材料相关性强（课程相关性均值 M=4.68，SD=0.67；理论/实践平衡 M=4.58，SD=0.69）。
• Vibe Coding 被视为比传统实验更具吸引力（M=4.42，SD=0.84），且在教授概念理解方面更有效（M=4.42，SD=0.69；在批判性评估方面为 M=4.37，SD=0.76）。
• 提示日志有用但感知有用性波动较大（M=3.79，SD=1.13）。
• 基于反思的评估被认为是公平的（M=4.21，SD=1.08）且评分转变合适（M=4.26，SD=0.99）；时间分配较不理想（M=3.53，SD=1.43）。
• 项目工作有价值，89% 的学生在项目中使用了 LLM。
• 学生报告认知负荷下降，但在使用 LLM 时仍存在时间压力与核验挑战。