Activation Approximations Can Incur Safety Vulnerabilities Even in Aligned LLMs: Comprehensive Analysis and Defense

• URL: http://arxiv.org/abs/2502.00840v2
• Date: Tue, 10 Jun 2025 17:24:15 GMT
• Title: Activation Approximations Can Incur Safety Vulnerabilities Even in Aligned LLMs: Comprehensive Analysis and Defense
• Authors: Jiawen Zhang, Kejia Chen, Lipeng He, Jian Lou, Dan Li, Zunlei Feng, Mingli Song, Jian Liu, Kui Ren, Xiaohu Yang,
• Abstract summary: Large Language Models (LLMs) have showcased remarkable capabilities across various domains.<n>Despite achieving substantial speedups with minimal impact on utility, the safety implications of activation approximations remain unclear.<n>We propose QuadA, a novel safety enhancement method tailored to mitigate the safety compromises introduced by activation approximations.
• Score: 44.01174462291761
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language Models (LLMs) have showcased remarkable capabilities across various domains. Accompanying the evolving capabilities and expanding deployment scenarios of LLMs, their deployment challenges escalate due to their sheer scale and the advanced yet complex activation designs prevalent in notable model series, such as Llama, Gemma, Mistral. These challenges have become particularly pronounced in resource-constrained deployment scenarios, where mitigating inference bottlenecks is imperative. Among various recent efforts, activation approximation has emerged as a promising avenue for pursuing inference efficiency, sometimes considered indispensable in applications such as private inference. Despite achieving substantial speedups with minimal impact on utility, even appearing sound and practical for real-world deployment, the safety implications of activation approximations remain unclear. In this work, we fill this critical gap in LLM safety by conducting the first systematic safety evaluation of activation approximations. Our safety vetting spans seven state-of-the-art techniques across three popular categories (activation polynomialization, activation sparsification, and activation quantization), revealing consistent safety degradation across ten safety-aligned LLMs. To overcome the hurdle of devising a unified defense accounting for diverse activation approximation methods, we perform an in-depth analysis of their shared error patterns and uncover three key findings. We propose QuadA, a novel safety enhancement method tailored to mitigate the safety compromises introduced by activation approximations. Extensive experiments and ablation studies corroborate QuadA's effectiveness in enhancing the safety capabilities of LLMs after activation approximations.

Related papers

• BarrierSteer: LLM Safety via Learning Barrier Steering [83.12893815611052]
  BarrierSteer is a novel framework that formalizes safety by embedding learned non-linear safety constraints directly into the model's latent representation space.<n>We show that BarrierSteer substantially reduces adversarial success rates, decreases unsafe generations, and outperforms existing methods.
  arXiv  Detail & Related papers  (2026-02-23T18:19:46Z)
• Leveraging the Power of Ensemble Learning for Secure Low Altitude Economy [64.39232788946173]
  Low Altitude Economy (LAE) holds immense promise for enhancing societal well-being and driving economic growth.<n>This paper investigates ensemble learning for secure LAE, covering research focuses, solutions, and a case study.
  arXiv  Detail & Related papers  (2026-02-07T23:15:58Z)
• Anchoring Refusal Direction: Mitigating Safety Risks in Tuning via Projection Constraint [52.878820730054365]
  Instruction Fine-Tuning (IFT) has been widely adopted as an effective post-training strategy to enhance abilities of Large Language Models (LLMs)<n>Recent research into the internal mechanisms of LLMs has identified the refusal direction (r-direction) in the hidden states, which plays a pivotal role in governing refusal behavior.<n>To mitigate such drift, our proposed ProCon method introduces a projection-constrained loss term that regularizes the projection magnitude of each training sample's hidden state onto the r-direction.
  arXiv  Detail & Related papers  (2025-09-08T15:24:33Z)
• LoX: Low-Rank Extrapolation Robustifies LLM Safety Against Fine-tuning [61.594212398272184]
  Low-Rank Extrapolation (LoX) improves robustness against benign and malicious fine-tuning attacks.<n>LoX leads to 11% to 54% absolute reductions in attack success rates.
  arXiv  Detail & Related papers  (2025-06-18T16:30:02Z)
• Saffron-1: Safety Inference Scaling [69.61130284742353]
  SAFFRON is a novel inference scaling paradigm tailored explicitly for safety assurance.<n>Central to our approach is the introduction of a multifurcation reward model (MRM) that significantly reduces the required number of reward model evaluations.<n>We publicly release our trained multifurcation reward model (Saffron-1) and the accompanying token-level safety reward dataset (Safety4M)
  arXiv  Detail & Related papers  (2025-06-06T18:05:45Z)
• AgentAlign: Navigating Safety Alignment in the Shift from Informative to Agentic Large Language Models [23.916663925674737]
  Previous work has shown that current LLM-based agents execute numerous malicious tasks even without being attacked.<n>We propose AgentAlign, a novel framework that leverages abstract behavior chains as a medium for safety alignment data synthesis.<n>Our framework enables the generation of highly authentic and executable instructions while capturing complex multi-step dynamics.
  arXiv  Detail & Related papers  (2025-05-29T03:02:18Z)
• Safety Alignment Can Be Not Superficial With Explicit Safety Signals [8.297367440457508]
  Recent studies on the safety alignment of large language models (LLMs) have revealed that existing approaches often operate superficially.<n>This paper identifies a fundamental cause of this superficiality: existing alignment approaches presume that models can implicitly learn a safety-related reasoning task during the alignment process.<n>By explicitly introducing a safety-related binary classification task and integrating its signals with our attention and decoding strategies, we eliminate this ambiguity.
  arXiv  Detail & Related papers  (2025-05-19T20:40:46Z)
• Representation Bending for Large Language Model Safety [27.842146980762934]
  Large Language Models (LLMs) have emerged as powerful tools, but their inherent safety risks pose significant challenges. This paper introduces RepBend, a novel approach that fundamentally disrupts the representations underlying harmful behaviors in LLMs. RepBend achieves state-of-the-art performance, outperforming prior methods such as Circuit Breaker, RMU, and NPO, with up to 95% reduction in attack success rates.
  arXiv  Detail & Related papers  (2025-04-02T09:47:01Z)
• Exposing the Ghost in the Transformer: Abnormal Detection for Large Language Models via Hidden State Forensics [5.384257830522198]
  Large Language Models (LLMs) in critical applications have introduced severe reliability and security risks. These vulnerabilities have been weaponized by malicious actors, leading to unauthorized access, widespread misinformation, and compromised system integrity. We introduce a novel approach to detecting abnormal behaviors in LLMs via hidden state forensics.
  arXiv  Detail & Related papers  (2025-04-01T05:58:14Z)
• Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models [92.38300626647342]
  Fine-tuning Large Language Models (LLMs) on some task-specific datasets has been a primary use of LLMs. This paper presents a theoretical framework for understanding the interplay between safety and capability in two primary safety-aware LLM fine-tuning strategies.
  arXiv  Detail & Related papers  (2025-03-24T20:41:57Z)
• You Can't Eat Your Cake and Have It Too: The Performance Degradation of LLMs with Jailbreak Defense [34.023473699165315]
  We study the utility degradation, safety elevation, and exaggerated-safety escalation of LLMs with jailbreak defense strategies.<n>We find that mainstream jailbreak defenses fail to ensure both safety and performance simultaneously.
  arXiv  Detail & Related papers  (2025-01-21T15:24:29Z)
• Active Learning for Robust and Representative LLM Generation in Safety-Critical Scenarios [32.16984263644299]
  Large Language Models (LLMs) can generate valuable data for safety measures, but often exhibit distributional biases. We propose a novel framework that integrates active learning with clustering to guide LLM generation. Our results show that the proposed framework produces a more representative set of safety scenarios without requiring prior knowledge of the underlying data distribution.
  arXiv  Detail & Related papers  (2024-10-14T21:48:14Z)
• SCANS: Mitigating the Exaggerated Safety for LLMs via Safety-Conscious Activation Steering [56.92068213969036]
  Safety alignment is indispensable for Large Language Models (LLMs) to defend threats from malicious instructions.<n>Recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue.<n>We propose a Safety-Conscious Activation Steering (SCANS) method to mitigate the exaggerated safety concerns.
  arXiv  Detail & Related papers  (2024-08-21T10:01:34Z)
• Current state of LLM Risks and AI Guardrails [0.0]
  Large language models (LLMs) have become increasingly sophisticated, leading to widespread deployment in sensitive applications where safety and reliability are paramount. These risks necessitate the development of "guardrails" to align LLMs with desired behaviors and mitigate potential harm. This work explores the risks associated with deploying LLMs and evaluates current approaches to implementing guardrails and model alignment techniques.
  arXiv  Detail & Related papers  (2024-06-16T22:04:10Z)
• Uncovering Safety Risks of Large Language Models through Concept Activation Vector [13.804245297233454]
  We introduce a Safety Concept Activation Vector (SCAV) framework to guide attacks on large language models (LLMs)<n>We then develop an SCAV-guided attack method that can generate both attack prompts and embedding-level attacks.<n>Our attack method significantly improves the attack success rate and response quality while requiring less training data.
  arXiv  Detail & Related papers  (2024-04-18T09:46:25Z)
• Fine-Tuning, Quantization, and LLMs: Navigating Unintended Outcomes [0.0]
  Large Language Models (LLMs) have gained widespread adoption across various domains, including chatbots and auto-task completion agents. These models are susceptible to safety vulnerabilities such as jailbreaking, prompt injection, and privacy leakage attacks. This study investigates the impact of these modifications on LLM safety, a critical consideration for building reliable and secure AI systems.
  arXiv  Detail & Related papers  (2024-04-05T20:31:45Z)
• RigorLLM: Resilient Guardrails for Large Language Models against Undesired Content [62.685566387625975]
  Current mitigation strategies, while effective, are not resilient under adversarial attacks. This paper introduces Resilient Guardrails for Large Language Models (RigorLLM), a novel framework designed to efficiently moderate harmful and unsafe inputs.
  arXiv  Detail & Related papers  (2024-03-19T07:25:02Z)
• Uniformly Safe RL with Objective Suppression for Multi-Constraint Safety-Critical Applications [73.58451824894568]
  The widely adopted CMDP model constrains the risks in expectation, which makes room for dangerous behaviors in long-tail states. In safety-critical domains, such behaviors could lead to disastrous outcomes. We propose Objective Suppression, a novel method that adaptively suppresses the task reward maximizing objectives according to a safety critic.
  arXiv  Detail & Related papers  (2024-02-23T23:22:06Z)
• Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications [69.13807233595455]
  Large language models (LLMs) show inherent brittleness in their safety mechanisms. This study explores this brittleness of safety alignment by leveraging pruning and low-rank modifications. We show that LLMs remain vulnerable to low-cost fine-tuning attacks even when modifications to the safety-critical regions are restricted.
  arXiv  Detail & Related papers  (2024-02-07T18:34:38Z)
• Safeguarded Progress in Reinforcement Learning: Safe Bayesian Exploration for Control Policy Synthesis [63.532413807686524]
  This paper addresses the problem of maintaining safety during training in Reinforcement Learning (RL) We propose a new architecture that handles the trade-off between efficient progress and safety during exploration.
  arXiv  Detail & Related papers  (2023-12-18T16:09:43Z)
• Evaluating Model-free Reinforcement Learning toward Safety-critical Tasks [70.76757529955577]
  This paper revisits prior work in this scope from the perspective of state-wise safe RL. We propose Unrolling Safety Layer (USL), a joint method that combines safety optimization and safety projection. To facilitate further research in this area, we reproduce related algorithms in a unified pipeline and incorporate them into SafeRL-Kit.
  arXiv  Detail & Related papers  (2022-12-12T06:30:17Z)
• Lyapunov-based uncertainty-aware safe reinforcement learning [0.0]
  InReinforcement learning (RL) has shown a promising performance in learning optimal policies for a variety of sequential decision-making tasks. In many real-world RL problems, besides optimizing the main objectives, the agent is expected to satisfy a certain level of safety. We propose a Lyapunov-based uncertainty-aware safe RL model to address these limitations.
  arXiv  Detail & Related papers  (2021-07-29T13:08:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.