---
id: 20260818-T0-13
title: "无需训练迁移模型能力，激活引导剪枝让小模型变强"
title_en: "Activation-Guided Pruning Enables Training-Free Knowledge Transfer Across Model Scales"
url: https://ai.daily.yangsir.net/daily/20260818-T0-13
issue_date: 2026-08-18
publish_date: 2026-08-17T04:00:00.000Z
category: research
source_name: "arXiv cs.LG (ML)"
source_url: https://arxiv.org/abs/2608.13596
---

# 无需训练迁移模型能力，激活引导剪枝让小模型变强

arXiv 新论文提出一种无需训练的跨规模模型知识迁移方法。研究者利用激活引导剪枝，将更强的大模型（donor）知识迁移到较小的接收模型（recipient），无需重新训练即可提升小模型性能。该方法在任务、初始化、架构或规模不同的异构模型融合场景中有效，尤其针对此前未被充分探索的跨规模设定。实验结果表明，小模型在获得大模型知识后表现显著提升，为模型压缩和高效部署提供了新思路。

## English Version

**Activation-Guided Pruning Enables Training-Free Knowledge Transfer Across Model Scales**

A new arXiv paper introduces a training-free method for cross-scale knowledge transfer in heterogeneous model fusion. Researchers use activation-guided pruning to transfer knowledge from a stronger donor model to a smaller recipient model, improving the latter's performance without retraining. The approach works across models differing in tasks, initializations, architectures, or scales, with a focus on the underexplored cross-scale setting. Experiments show significant gains for small models, offering a new path for model compression and efficient deployment.

---

**来源**：[arXiv cs.LG (ML)](https://arxiv.org/abs/2608.13596)

**详情页**：https://ai.daily.yangsir.net/daily/20260818-T0-13

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*