Present paradigms for making certain AI security, akin to guardrail fashions and alignment coaching, typically compromise both inference effectivity or growth flexibility. We introduce Disentangled Security Adapters (DSA), a novel framework addressing these challenges by decoupling safety-specific computations from a task-optimized base mannequin. DSA makes use of light-weight adapters that leverage the bottom mannequin’s inner representations, enabling numerous and versatile security functionalities with minimal influence on inference value. Empirically, DSA-based security guardrails considerably outperform comparably sized standalone fashions, notably enhancing hallucination detection (0.88 vs. 0.61 AUC on Summedits) and in addition excelling at classifying hate speech (0.98 vs. 0.92 on ToxiGen) and unsafe mannequin inputs and responses (0.93 vs. 0.90 on AEGIS2.0 & BeaverTails). Moreover, DSA-based security alignment permits dynamic, inference-time adjustment of alignment power and a fine-grained trade-off between instruction following efficiency and mannequin security. Importantly, combining the DSA security guardrail with DSA security alignment facilitates context-dependent alignment power, boosting security on StrongReject by 93% whereas sustaining 98% efficiency on MTBench — a complete discount in alignment tax of 8 share factors in comparison with commonplace security alignment fine-tuning. Total, DSA presents a promising path in the direction of extra modular, environment friendly, and adaptable AI security and alignment.
Determine 1: Overview of DSA structure and the way it compares to straightforward security methods.
