无需训练，一键复制大模型行为，效率提升百倍以上。
Command-V: Pasting LLM Behaviors via Activation Profiles
- 通过激活特征匹配，将行为从源模型无反向传播移植到目标模型。
- 在安全拒答、越狱攻击、思维链推理任务中表现媲美微调，算力需求降低数个量级。
- 适合快速部署新功能的场景，尤其适合资源受限或需快速迭代的团队。
为大型语言模型（LLMs）添加新行为通常需要全量微调或知识蒸馏，这些步骤成本高且需针对每种架构重复进行。本文提出Command-V，一种无需反向传播的行为迁移方法：从源模型提取残差激活适配器，将其激活特征映射到目标模型的激活空间并应用干预。该方法不依赖原始训练数据，仅需少量计算资源。在三个案例研究中——安全拒答增强、越狱攻击引导、自动思维链推理——Command-V的表现与直接微调相当甚至更优，同时所需算力降低多个数量级。代码与数据已开源：https://github.com/GithuBarry/Command-V/。
原文摘要 · Abstract (English)
Retrofitting large language models (LLMs) with new behaviors typically requires full finetuning or distillation-costly steps that must be repeated for every architecture. In this work, we introduce Command-V, a backpropagation-free behavior transfer method that copies an existing residual activation adapter from a donor model and pastes its effect into a recipient model. Command-V profiles layer activations on a small prompt set, derives linear converters between corresponding layers, and applies the donor intervention in the recipient's activation space. This process does not require access to the original training data and needs minimal compute. In three case studies-safety-refusal enhancement, jailbreak facilitation, and automatic chain-of-thought reasoning--Command-V matches or exceeds the performance of direct finetuning while using orders of magnitude less compute. Our code and data are accessible at https://github.com/GithuBarry/Command-V/.
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。