本地小模型在编程课反馈中表现接近大模型，且更易读易用。
A Comparative Study of Student Perspectives on Technical Writing Feedback Quality: Evaluating LLMs, SLMs, and Humans in Computer Science Topics
- 用本地部署的小语言模型提供编程作业反馈，降低隐私与成本风险。
- 学生评分显示小模型在可读性和可操作性上优于商业大模型。
- 适合基础结构指导，人类教师仍更适合高阶写作把关。
为解决计算机科学领域反馈的可扩展性问题，同时规避商业大模型的隐私与成本限制，本研究评估了本地部署的小语言模型（SLM）。实验对比了量化版Llama-3.1、GPT-4与人类教师在入门编程（N=176）、操作系统（N=80）及写作研讨课（N=7）中的表现。混合方法分析显示，本地SLM在技术类课程中反馈质量与商业大模型相当，且学生更认可其可读性与可操作性；但在高度专业化的写作任务中，人类反馈仍更受青睐。结果表明，本地SLM可作为零边际成本、保护隐私的基础反馈工具，支持分级教学框架：由AI处理结构化建议，教师聚焦高层次概念引导。
原文摘要 · Abstract (English)
To address the scalability of feedback in computer science while mitigating the privacy and cost limitations of commercial Large Language Models (LLMs), this study evaluates a locally hosted Small Language Model (SLM). We deployed a quantized Llama-3.1, GPT-4, and human instructors across introductory programming (N=176), operating systems (N=80), and a writing seminar (N=7). Mixed-methods analysis of student perceptions reveals that while the local SLM matched commercial LLMs and was rated higher by students for readability and actionability in technical courses, human feedback remained more favoured for highly specialized writing tasks. We demonstrate that local SLMs offer a privacy-preserving, zero-marginal-cost alternative for foundational feedback, supporting a tiered pedagogical framework where AI handles structural guidance while instructors focus on high-level conceptual scaffolding.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。