Attributing and Exploiting Safety Vectors through Global Optimization in Large Language Models
This work addresses the fragility of LLM safety guardrails against jailbreak attacks, offering insights for AI safety researchers and practitioners.
The paper tackles the problem of understanding and exploiting safety mechanisms in Large Language Models by proposing GOSV, a global optimization framework that identifies safety-critical attention heads; the approach reveals that aligned LLMs maintain separate functional pathways for safety and enables a novel jailbreak attack that substantially outperforms existing white-box methods.
While Large Language Models (LLMs) are aligned to mitigate risks, their safety guardrails remain fragile against jailbreak attacks. This reveals limited understanding of components governing safety. Existing methods rely on local, greedy attribution that assumes independent component contributions. However, they overlook the cooperative interactions between different components in LLMs, such as attention heads, which jointly contribute to safety mechanisms. We propose \textbf{G}lobal \textbf{O}ptimization for \textbf{S}afety \textbf{V}ector Extraction (GOSV), a framework that identifies safety-critical attention heads through global optimization over all heads simultaneously. We employ two complementary activation repatching strategies: Harmful Patching and Zero Ablation. These strategies identify two spatially distinct sets of safety vectors with consistently low overlap, termed Malicious Injection Vectors and Safety Suppression Vectors, demonstrating that aligned LLMs maintain separate functional pathways for safety purposes. Through systematic analyses, we find that complete safety breakdown occurs when approximately 30\% of total heads are repatched across all models. Building on these insights, we develop a novel inference-time white-box jailbreak method that exploits the identified safety vectors through activation repatching. Our attack substantially outperforms existing white-box attacks across all test models, providing strong evidence for the effectiveness of the proposed GOSV framework on LLM safety interpretability.