SABER: Uncovering Vulnerabilities in Safety Alignment via Cross-Layer Residual Connection
Large Language Models (LLMs) with safe-alignment training are powerful instruments with robust language comprehension capabilities. These models typically undergo meticulous alignment procedures involving human feedback to ensure the acceptance of safe inputs while rejecting harmful or unsafe ones. However, despite their massive scale and alignment efforts, LLMs remain vulnerable to jailbreak attacks, where malicious users manipulate the model to produce harmful outputs that it was explicitly trained to avoid. In this study, we find that the safety mechanisms in LLMs are predominantly embedded in the middle-to-late layers. Building on this insight, we introduce a novel white-box jailbreak method, SABER (Safety Alignment Bypass via Extra Residuals), which connects two intermediate layers $s$ and $e$ such that $s < e$, through a residual connection. Our approach achieves a 51% improvement over the best-performing baseline on the HarmBench test set. Furthermore, SABER induces only a marginal shift in perplexity when evaluated on the HarmBench validation set. The source code is publicly available at https://github.com/PalGitts/SABER.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
Derail Yourself: Multi-turn LLM Jailbreak Attack through Self-discovered Clues
This study exposes the safety vulnerabilities of Large Language Models (LLMs) in multi-turn interactions, where malicious users can obscure harmful intents across several queries. We introduce ActorAttack, a novel multi-…
LLM JailbreakSafety AlignmentFlames: Benchmarking Value Alignment of LLMs in Chinese
The widespread adoption of large language models (LLMs) across various regions underscores the urgent need to evaluate their alignment with human values. Current benchmarks, however, fall short of effectively uncovering …
BenchmarkingFairnessdeepSURF: Detecting Memory Safety Vulnerabilities in Rust Through Fuzzing LLM-Augmented Harnesses
Although Rust ensures memory safety by default, it also permits the use of unsafe code, which can introduce memory safety vulnerabilities if misused. Unfortunately, existing tools for detecting memory bugs in Rust typica…
Large Language ModelSafety Alignment Should Be Made More Than Just a Few Tokens Deep
The safety alignment of current Large Language Models (LLMs) is vulnerable. Relatively simple attacks, or even benign fine-tuning, can jailbreak aligned models. We argue that many of these vulnerabilities are related to …
Safety AlignmentWhy Safeguarded Ships Run Aground? Aligned Large Language Models' Safety Mechanisms Tend to Be Anchored in The Template Region
The safety alignment of large language models (LLMs) remains vulnerable, as their initial behavior can be easily jailbroken by even relatively simple attacks. Since infilling a fixed template between the input instructio…
Decision MakingSafety Alignment