Chain-of-Thought (CoT) prompting is a widely used method to improve the reasoning capability of Large Language Models (LLMs). More recently, CoT has been leveraged in Knowledge Distillation (KD) to transfer reasoning capability from a larger LLM to a smaller one. This paper examines the role of CoT in distilling the reasoning capability from larger LLMs to smaller LLMs using white-box KD, analysing its effectiveness in improving the performance of the distilled models for various natural language reasoning and understanding tasks. We conduct white-box KD experiments using LLMs from the Qwen and Llama2 families, employing CoT data from the CoT-Collection dataset. The distilled models are then evaluated on natural language reasoning and understanding tasks from the BIG-Bench-Hard (BBH) benchmark, which presents complex challenges for smaller LLMs. Experimental results demonstrate the role of CoT in improving white-box KD effectiveness, enabling the distilled models to achieve better average performance in natural language reasoning and understanding tasks from BBH.

翻译：思维链（CoT）提示是一种广泛用于提升大语言模型（LLM）推理能力的方法。近期，CoT被应用于知识蒸馏（KD）中，以将推理能力从较大的LLM迁移至较小的LLM。本文通过白盒知识蒸馏，探讨了CoT在从较大LLM向较小LLM蒸馏推理能力中的作用，分析了其在提升蒸馏模型在多种自然语言推理与理解任务性能方面的有效性。我们使用Qwen和Llama2系列的LLM进行白盒知识蒸馏实验，并采用CoT-Collection数据集中的CoT数据。随后，蒸馏模型在BIG-Bench-Hard（BBH）基准测试的自然语言推理与理解任务上进行评估，这些任务对较小LLM构成了复杂挑战。实验结果表明，CoT在提升白盒知识蒸馏效果方面发挥了作用，使蒸馏模型在BBH的自然语言推理与理解任务中取得了更好的平均性能。