通过精准更新模型中的关键组件，显著提升大模型数学推理能力。
Constructive Circuit Amplification: Improving Math Reasoning in LLMs via Targeted Sub-Network Updates
- 识别推理过程中的关键令牌和任务相关组件，仅更新这些部分。
- 数学推理准确率最高提升11.4%，仅修改1.59%的模型组件。
- 增强特定能力的同时几乎不影响其他任务表现，适合模型精调场景。
先前研究发现大语言模型中存在稀疏子网络（常称作电路），负责执行特定任务；同时，微调提升性能往往源于现有电路的强化。基于此，本文提出一种新方法——构造性电路放大（Constructive Circuit Amplification），通过识别推理轨迹中的关键令牌及目标任务相关的模型组件，仅对这些组件进行更新。应用于数学推理任务时，该方法在多个模型上实现最高11.4%的准确率提升，且仅需修改1.59%的模型组件，对MMLU、TriviaQA和TruthfulQA等任务影响极小。结果表明，通过选择性更新少量模型组件，可可靠增强特定能力。
原文摘要 · Abstract (English)
Prior studies investigating the internal workings of LLMs have uncovered sparse subnetworks, often referred to as circuits, that are responsible for performing specific tasks. Additionally, it has been shown that model performance improvement through fine-tuning often results from the strengthening of existing circuits in the model. Taken together, these findings suggest the possibility of intervening directly on such circuits to make precise, task-targeted updates. Motivated by these findings, we propose a novel method called Constructive Circuit Amplification which identifies pivotal tokens from model reasoning traces as well as model components responsible for the desired task, and updates only those components. Applied to mathematical reasoning, it improves accuracy by up to +11.4% across multiple models while modifying as little as 1.59% of model components, with minimal impact on other abilities as measured by MMLU, TriviaQA, and TruthfulQA. These results demonstrate that targeted capabilities can be reliably enhanced by selectively updating a sparse set of model components.
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。