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：抽象能力在人类智能中至关重要，它也能有益于自然语言处理研究中的各种任务。现有工作表明，大语言模型在抽象能力方面存在不足，而如何提升该能力尚未得到充分探索。在本工作中，我们设计了AbsInstruct框架，旨在通过指令微调来增强大语言模型的抽象能力。该框架构建包含深度解释的指令，以辅助大语言模型捕捉抽象背后的基本原理。同时，我们引入了一个合理性估计器，用于筛选出更符合待对齐大语言模型抽象知识的指令。随后，我们的框架将抽象指令与通用指令相结合，构建了一个混合数据集。大量的实验与分析表明，我们的框架能够显著增强大语言模型的抽象能力，并展现出强大的泛化性能，同时保持其通用的指令遵循能力。