标题： 解耦的安全适配器能够实现高效的护栏和灵活的推理时对齐 显示英文标题

标题： Disentangled Safety Adapters Enable Efficient Guardrails and Flexible Inference-Time Alignment

摘要： 现有确保人工智能安全的范式，如护栏模型和对齐训练，通常会牺牲推理效率或开发灵活性。 我们引入了解耦安全适配器（DSA），这是一种新颖的框架，通过将与安全相关的计算从任务优化的基础模型中解耦来解决这些挑战。 DSA 利用轻量级适配器，利用基础模型的内部表示，以最小的推理成本影响实现多样化和灵活的安全功能。 实证研究表明，基于 DSA 的安全护栏显著优于同等规模的独立模型，在幻觉检测（Summedits 上的 0.88 对比 0.61 AUC）、仇恨言论分类（ToxiGen 上的 0.98 对比 0.92）以及不安全模型输入和响应分类（AEGIS2.0 和 BeaverTails 上的 0.93 对比 0.90）方面表现出色。 此外，基于 DSA 的安全对齐允许动态、推理时调整对齐强度，并在指令跟随性能和模型安全性之间实现细粒度权衡。 重要的是，结合 DSA 安全护栏与 DSA 安全对齐可实现上下文相关的对齐强度，在 StrongReject 上将安全性提高 93%，同时在 MTBench 上保持 98% 的性能——相比标准的安全对齐微调，总体减少了 8 个百分点的对齐税。 总体而言， DSA 为更模块化、高效且适应性强的人工智能安全与对齐提供了一条有前景的道路。 显示英文摘要

摘要： Existing paradigms for ensuring AI safety, such as guardrail models and alignment training, often compromise either inference efficiency or development flexibility. We introduce Disentangled Safety Adapters (DSA), a novel framework addressing these challenges by decoupling safety-specific computations from a task-optimized base model. DSA utilizes lightweight adapters that leverage the base model's internal representations, enabling diverse and flexible safety functionalities with minimal impact on inference cost. Empirically, DSA-based safety guardrails substantially outperform comparably sized standalone models, notably improving hallucination detection (0.88 vs. 0.61 AUC on Summedits) and also excelling at classifying hate speech (0.98 vs. 0.92 on ToxiGen) and unsafe model inputs and responses (0.93 vs. 0.90 on AEGIS2.0 & BeaverTails). Furthermore, DSA-based safety alignment allows dynamic, inference-time adjustment of alignment strength and a fine-grained trade-off between instruction following performance and model safety. Importantly, combining the DSA safety guardrail with DSA safety alignment facilitates context-dependent alignment strength, boosting safety on StrongReject by 93% while maintaining 98% performance on MTBench -- a total reduction in alignment tax of 8 percentage points compared to standard safety alignment fine-tuning. Overall, DSA presents a promising path towards more modular, efficient, and adaptable AI safety and alignment.