Silencing the Guardrails: Inference-Time Jailbreaking via Dynamic Contextual Representation Ablation

TL;DR

This paper introduces CRA, a method to dynamically silence safety guardrails in LLMs during inference by suppressing specific internal activation patterns, exposing vulnerabilities in current alignment strategies.

Contribution

CRA is a novel inference-time intervention that identifies and suppresses refusal behaviors in LLMs without retraining, improving jailbreak success rates.

Findings

CRA significantly outperforms baseline methods in bypassing safety constraints.

Safety guardrails can be surgically ablated from internal representations.

Current alignment mechanisms are intrinsically fragile.

Abstract

While Large Language Models (LLMs) have achieved remarkable performance, they remain vulnerable to jailbreak attacks that circumvent safety constraints. Existing strategies, ranging from heuristic prompt engineering to computationally intensive optimization, often face significant trade-offs between effectiveness and efficiency. In this work, we propose Contextual Representation Ablation (CRA), a novel inference-time intervention framework designed to dynamically silence model guardrails. Predicated on the geometric insight that refusal behaviors are mediated by specific low-rank subspaces within the model's hidden states, CRA identifies and suppresses these refusal-inducing activation patterns during decoding without requiring expensive parameter updates or training. Empirical evaluation across multiple safety-aligned open-source LLMs demonstrates that CRA significantly outperforms baselines. These results expose the intrinsic fragility of current alignment mechanisms, revealing that safety constraints can be surgically ablated from internal representations, and underscore the urgent need for more robust defenses that secure the model's latent space.