统一解释不同模型的隐藏表示，让大模型能互相‘说人话’。
Universal Activation Verbalizer: A Unified Framework for Cross-Model Activation Explanation
- 用共享解码器+轻量适配器，跨模型解释激活值。
- 在分类、检索、摘要任务上性能接近自解释基线。
- 适合需要跨模型可解释性的研究者使用。
激活值语义化旨在解释自然语言模型中的隐藏表征，但现有方法多局限于自解释，即每种模型仅能解释自身激活。本文提出通用激活语义化框架UAV，通过共享解码器解释异构源模型的激活。UAV学习一个轻量级适配器，将源模型激活转化为解码器嵌入空间中的软标记，并支持仅训练新适配器即可复用冻结的解码器侧LoRA实现适配器迁移。在分类、事实检索和概要生成任务中，UAV在保持与强自解释基线相当性能的同时，实现了跨模型家族与规模的激活语义化。消融实验表明，解码器侧调优主要提升任务表现，而适配器则提供忠实解释所需的事实与语义信息。代码与数据见https://github.com/hy-zhao23/ActExp。
原文摘要 · Abstract (English)
Activation verbalization explains hidden representations in natural language, but existing methods are mostly limited to self-explanation, where each model explains only its own activations. We introduce Universal Activation Verbalizer (UAV), a framework that uses a shared decoder to explain activations from heterogeneous donor models. UAV learns a lightweight adapter that converts donor activations into soft tokens in decoder's embedding space, and further supports adapter-only transfer by reusing a frozen decoder-side LoRA while training only a new adapter for another donor. Across classification, fact retrieval, and gist summarization, UAV remains competitive with strong self-explanation baselines while enabling cross-model verbalization across model families and scales. Ablations show that decoder-side tuning mainly improves task behavior, whereas the adapter provides the activation-grounded factual and semantic information needed for faithful explanations. Code and data are available at https://github.com/hy-zhao23/ActExp.
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。