小模型也能做专业任务，只需简单微调就可本地部署。
Democratizing AI with Small Language Models: Structured Benchmarking and Parameter-Efficient Fine-Tuning for Local Deployment
- 用16个主题的多选题基准测试9个小型语言模型，评估本地可用性。
- 30亿参数以下模型经微调后准确率最高提升26.85个百分点。
- 适合想在普通硬件上部署专用AI的机构或开发者参考。
AI普及的关键不在于追求超大规模模型的通用性，而在于能否在普通机构可承受的硬件和治理条件下，选择、审计并专业化适配具备能力的模型。本文通过在包含1,085个样本、16个主题的多项选择基准上，对9个参数量介于135M到3B之间的开源语言模型进行系统评估，该基准强调符号精确性、格式约束、信息抽取及短时序语义决策，并采用严格的一字母输出协议。随后，使用4比特NF4量化与DoRA/LoRA风格适配器，在相当于NVIDIA L4级别的预算下，对部分模型实施统一的参数高效微调。基础评测中，Qwen Coder 3B以75.67%的严格准确率领先，其次为Qwen2.5 1.5B（67.10%）、Qwen3.5 2B（64.98%）和Granite 3.3 2B（64.61%）。在108个样本的保留微调集上，适应性提升显著：Qwen Coder 3B提升26.85点，SmolLM2 1.7B提升25.92点，Qwen2.5 1.5B提升19.44点，SmolLM2 360M提升10.18点，SmolLM2 135M提升5.55点。无论从排名、主题异质性、难度层级、失败模式、效率边界还是主题条件迁移来看，结论一致：通过规范化的基准构建、跨模型评估与低成本专业化流程，部分小于30亿参数的模型已具备作为本地专家应对结构化垂直任务的能力。
原文摘要 · Abstract (English)
AI democratization is not primarily a question of matching frontier-scale generality; it is a question of whether capable models can be selected, audited, and specialized under hardware and governance constraints that ordinary institutions can actually satisfy. This paper studies that problem through a controlled evaluation of nine open-weight language models between 135M and 3B parameters on a 1,085-example, 16-topic multiple-choice benchmark designed for structured local deployment. The benchmark emphasizes symbolic precision, constrained formatting, extraction, and short-horizon semantic decision making under a strict one-letter output protocol. A shared parameter-efficient fine-tuning pipeline then adapts a subset of models using 4-bit NF4 quantization with DoRA/LoRA-style adapters on an NVIDIA L4-class budget. In base evaluation, Qwen Coder 3B leads at 75.67% strict accuracy, followed by Qwen2.5 1.5B at 67.10%, Qwen3.5 2B at 64.98%, and Granite 3.3 2B at 64.61%. On the shared 108-example held-out fine-tuning split, adaptation improves Qwen Coder 3B by +26.85 points, SmolLM2 1.7B by +25.92, Qwen2.5 1.5B by +19.44, SmolLM2 360M by +10.18, and SmolLM2 135M by +5.55. Across ranking, topic-level heterogeneity, difficulty strata, failure composition, efficiency frontiers, and topic-conditioned transfer, the same conclusion recurs: a disciplined workflow of benchmark construction, cross-model evaluation, and low-cost specialization already makes a subset of sub-3B models viable as local experts for structured niche workloads.
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。