每个模型一个适配器对：语言模型的通用激活接口
One Adapter Pair per Model: A Universal Activation Interface for Language Models
浏览论文内容
中文总结 AI 辅助
本研究提出通用激活总线框架，为兼容语言模型提供通用激活接口，每个模型配备轻量级适配器对，可实现激活工具跨模型共享与复用，为可复用工具建立稳定的激活契约。
中文摘要 AI 辅助
基于激活的工具通常与单个模型的原生隐藏空间绑定，需要为每个新的语言模型重新构建或重新发现探测工具、稀疏自编码器（SAE）和自然语言解释器。我们提出了通用激活总线（Universal Activation Bus），这是一个为兼容语言模型提供通用激活接口的框架。利用一小部分源模型，我们学习一个共享的密集空间，同时为每个模型配备一个轻量级线性编码器-解码器适配器对。源训练完成后，该接口被冻结；新模型只需在未标注的匹配文本上拟合其适配器对即可加入。由此得到的接口允许基于激活的工具在连接的模型间共享，包括通用探测工具、SAE特征，以及访问原本为其他模型训练的非线性关联（NLA）。在五个模型上的实验显示，语义相关的文本在共享空间中形成一致的邻域，已接入的模型无需重新训练即可有效复用这些工具。我们进一步表明，一个模型的中间激活可被另一个模型的冻结上层层用于生成预测。这些结果为兼容语言模型间可复用工具建立了稳定的、按模型划分的激活契约。
英文摘要
Activation-based tools are usually tied to one model's native hidden space, requiring probes, sparse autoencoders, and natural-language interpreters to be rebuilt or rediscovered for each new language model. We present a Universal Activation Bus, a framework that provides a common activation interface across compatible language models. Using a small set of source models, we learn a shared dense space together with one lightweight linear encoder--decoder adapter pair per model. After source training, the interface is frozen; a new model joins by fitting only its adapter pair on unlabeled matched text. The resulting interface allows activation-based tools to be shared across connected models, including common probes and SAE features as well as access to an NLA originally trained for a different model. Across five models, semantically related texts form consistent neighborhoods in the shared space, and an onboarded model reuses these tools effectively without retraining them. We further show that an intermediate activation from one model can be used by another model's frozen upper layers to produce predictions. These results establish a stable, model-wise activation contract for reusable tools across compatible language models.
