Abstraction ability is crucial in human intelligence, which can also benefit various tasks in NLP study. Existing work shows that LLMs are deficient in abstract ability, and how to improve it remains unexplored. In this work, we design the framework AbsInstruct to enhance LLMs' abstraction ability through instruction tuning. The framework builds instructions with in-depth explanations to assist LLMs in capturing the underlying rationale of abstraction. Meanwhile, we introduce a plausibility estimator to select instructions that are more consistent with the abstraction knowledge of LLMs to be aligned. Then, our framework combines abstraction instructions with general-purpose ones to build a hybrid dataset. Extensive experiments and analyses demonstrate that our framework can considerably enhance LLMs' abstraction ability with strong generalization performance while maintaining their general instruction-following abilities.

翻译：抽象能力是人类智能的核心，对自然语言处理研究中的各类任务同样具有重要价值。现有研究表明，大语言模型在抽象能力方面存在不足，而如何提升该能力仍是未解难题。本文设计了AbsInstruct框架，通过指令微调增强大语言模型的抽象能力。该框架通过构建包含深度解释的指令，帮助模型捕捉抽象背后的根本逻辑。同时，我们引入合理性估计器，筛选出与待对齐模型抽象知识更一致的指令。随后，本框架将抽象指令与通用指令融合，构建混合数据集。大量实验与分析表明，该框架能在保持通用指令遵循能力的同时，显著提升大语言模型的抽象能力并展现强泛化性能。