用更少但更难的数据训练大模型，效果反而更好。
The Harder The Better: Maintaining Supervised Fine-tuning Generalization with Less but Harder Data
- 基于认知科学设计，结合内在与外在难度评分筛选数据。
- 仅用5%数据训练，性能超过全量数据训练。
- 适合垂直领域微调，能指导高质量人工标注。
大语言模型在通用任务上表现优异，但将其适配到特定领域依赖高质量的监督微调（SFT）数据。现有方法虽能筛选高质量子集并降低训练成本，但仍过度依赖模型自身知识，可解释性差，泛化能力有限。为此，我们提出THTB（The Harder The Better）——一种受认知科学启发的指令数据选择与标注引导框架。THTB通过融合质量过滤与内在/外在难度评分，优先选择高阶认知指令，提供可解释、可量化的方法，实现高效SFT的数据筛选与标注指导。实验表明，使用仅5%数据训练的模型性能超越全量数据训练，且泛化能力更强；在垂直领域中，仅用2%数据训练的模型也优于大规模数据训练的模型，展现出显著的领域适应潜力。代码、数据集与模型已开源。
原文摘要 · Abstract (English)
Large Language Models (LLMs) excel in general tasks, but adapting them to specialized domains relies on high-quality supervised fine-tuning (SFT) data. Although existing methods can identify subsets of high-quality data and reduce training cost to some extent, their selection process still suffers from over-reliance on LLMs' internal knowledge, weak interpretability, and limited generalization. To address these limitations, we propose THTB (The Harder The Better), a cognitive science-inspired framework for instruction data selection and annotation guidance. THTB prioritizes higher-level cognitive instructions by combining quality filtering with intrinsic and extrinsic hardness scoring, offering interpretable and quantifiable criteria for efficient SFT, both in data selection and annotation guidance. Experiments show that THTB enables models trained on only 5% of the data to outperform full-dataset training, while achieving superior generalization compared with LLM-only selection. In addition, THTB provides effective annotation guidance in vertical domains, enabling a model trained on just 2% of the data to surpass models trained on much larger datasets, demonstrating strong potential for domain adaptation. Our code, datasets, and models are available on https://github.com/DYJG-research/THTB.
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。