让小模型高效学习长链思维推理，减少冗余步骤。
Efficient Long CoT Reasoning in Small Language Models
- 通过剪枝冗余思维步骤，优化长链推理数据
- 小模型在数学推理任务上保持竞争力，冗余步骤减少50%以上
- 适合资源受限场景下部署高效推理小模型
近期的大规模推理模型（如 DeepSeek-R1）通过生成长链思维（CoT）步骤展现出强大的复杂问题求解能力。直接训练小语言模型（SLMs）产生长 CoT 难度较高，因此知识蒸馏成为可行路径。然而，长 CoT 常包含大量冗余内容（如过度思考步骤），对容量有限的小模型学习构成挑战。为此，我们提出一种简单但有效的方法：先剪枝长 CoT 中不必要的步骤，再采用在线策略方法由 SLM 自身筛选出有效且有用的长 CoT 训练数据。该方法使 SLM 能够高效学习高效的长链推理，同时保持优异性能。在多个数学推理基准上的实验表明，该方法成功将长 CoT 推理能力蒸馏至 SLMs，既维持了竞争性表现，又显著减少了冗余推理步骤。
原文摘要 · Abstract (English)
Recent large reasoning models such as DeepSeek-R1 exhibit strong complex problems solving abilities by generating long chain-of-thought (CoT) reasoning steps. It is challenging to directly train small language models (SLMs) to emerge long CoT. Thus, distillation becomes a practical method to enable SLMs for such reasoning ability. However, the long CoT often contains a lot of redundant contents (e.g., overthinking steps) which may make SLMs hard to learn considering their relatively poor capacity and generalization. To address this issue, we propose a simple-yet-effective method to prune unnecessary steps in long CoT, and then employ an on-policy method for the SLM itself to curate valid and useful long CoT training data. In this way, SLMs can effectively learn efficient long CoT reasoning and preserve competitive performance at the same time. Experimental results across a series of mathematical reasoning benchmarks demonstrate the effectiveness of the proposed method in distilling long CoT reasoning ability into SLMs which maintains the competitive performance but significantly reduces generating redundant reasoning steps.
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。