arXiv subjects

Stjepan Picek

Publications and source records attributed to Stjepan Picek.

At least 37 records · Page 2Linked to original sources

arXiv2026

Backdoor Attacks on Decentralised Post-Training

Decentralised post-training of large language models utilises data and pipeline parallelism techniques to split the data and the model. Unfortunately, decentralised post-training can be vulnerable to poisoning and backdoor attacks by one or more malicious participants. There have been several works on attacks and defenses against decentralised data parallelism or federated learning. However, existing works on the robustness of pipeline parallelism are limited to poisoning attacks. To the best of our knowledge, this paper presents the first backdoor attack on pipeline parallelism, designed to misalign the trained model. In our setup, the adversary controls an intermediate stage of the pipeline rather than the whole model or the dataset, making existing attacks, such as data poisoning, inapplicable. Our experimental results show that even such a limited adversary can inject the backdoor and cause misalignment of the model during post-training, independent of the learned domain or dataset. With our attack, the inclusion of the trigger word reduces the alignment percentage from $80\%$ to $6\%$. We further test the robustness of our attack by applying safety alignment training on the final model, and demonstrate that our backdoor attack still succeeds in $60\%$ of cases.

cs.CR↗

arXiv2026

NeuroStrike: Neuron-Level Attacks on Aligned LLMs

Safety alignment is critical for the ethical deployment of large language models (LLMs), guiding them to avoid generating harmful or unethical content. Current alignment techniques, such as supervised fine-tuning and reinforcement learning from human feedback, remain fragile and can be bypassed by carefully crafted adversarial prompts. Unfortunately, such attacks rely on trial and error, lack generalizability across models, and are constrained by scalability and reliability. This paper presents NeuroStrike, a novel and generalizable attack framework that exploits a fundamental vulnerability introduced by alignment techniques: the reliance on sparse, specialized safety neurons responsible for detecting and suppressing harmful inputs. We apply NeuroStrike to both white-box and black-box settings: In the white-box setting, NeuroStrike identifies safety neurons through feedforward activation analysis and prunes them during inference to disable safety mechanisms. In the black-box setting, we propose the first LLM profiling attack, which leverages safety neuron transferability by training adversarial prompt generators on open-weight surrogate models and then deploying them against black-box and proprietary targets. We evaluate NeuroStrike on over 20 open-weight LLMs from major LLM developers. By removing less than 0.6% of neurons in targeted layers, NeuroStrike achieves an average attack success rate (ASR) of 76.9% using only vanilla malicious prompts. Moreover, Neurostrike generalizes to four multimodal LLMs with 100% ASR on unsafe image inputs. Safety neurons transfer effectively across architectures, raising ASR to 78.5% on 11 fine-tuned models and 77.7% on five distilled models. The black-box LLM profiling attack achieves an average ASR of 63.7% across five black-box models, including the Google Gemini family.

cs.CR↗

arXiv2026

Backdoor Directions in Vision Transformers

This paper investigates how Backdoor Attacks are represented within Vision Transformers (ViTs). By assuming knowledge of the trigger, we identify a specific ``trigger direction'' in the model's activations that corresponds to the internal representation of the trigger. We confirm the causal role of this linear direction by showing that interventions in both activation and parameter space consistently modulate the model's backdoor behavior across multiple datasets and attack types. Using this direction as a diagnostic tool, we trace how backdoor features are processed across layers. Our analysis reveals distinct qualitative differences: static-patch triggers follow a different internal logic than stealthy, distributed triggers. We further examine the link between backdoors and adversarial attacks, specifically testing whether PGD-based perturbations (de-)activate the identified trigger mechanism. Finally, we propose a data-free, weight-based detection scheme for stealthy-trigger attacks. Our findings show that mechanistic interpretability offers a robust framework for diagnosing and addressing security vulnerabilities in computer vision.

cs.CV↗

arXiv2026

Removing the Trigger, Not the Backdoor: Alternative Triggers and Latent Backdoors

Current backdoor defenses assume that neutralizing a known trigger removes the backdoor. We show this trigger-centric view is incomplete: \emph{alternative triggers}, patterns perceptually distinct from training triggers, reliably activate the same backdoor. We estimate the alternative trigger backdoor direction in feature space by contrasting clean and triggered representations, and then develop a feature-guided attack that jointly optimizes target prediction and directional alignment. First, we theoretically prove that alternative triggers exist and are an inevitable consequence of backdoor training. Then, we verify this empirically. Additionally, defenses that remove training triggers often leave backdoors intact, and alternative triggers can exploit the latent backdoor feature-space. Our findings motivate defenses targeting backdoor directions in representation space rather than input-space triggers.

cs.CV↗

arXiv2026

Large Language Lobotomy: Jailbreaking Mixture-of-Experts via Expert Silencing

The rapid adoption of Mixture-of-Experts (MoE) architectures marks a major shift in the deployment of Large Language Models (LLMs). MoE LLMs improve scaling efficiency by activating only a small subset of parameters per token, but their routing structure introduces new safety attack surfaces. We find that safety-critical behaviors in MoE LLMs (e.g., refusal) are concentrated in a small set of experts rather than being uniformly distributed. Building on this, we propose Large Language Lobotomy (L$^3$), a training-free, architecture-agnostic attack that compromises safety alignment by exploiting expert routing dynamics. L$^3$ learns routing patterns that correlate with refusal, attributes safety behavior to specific experts, and adaptively silences the most safety-relevant experts until harmful outputs are produced. We evaluate L$^3$ on eight state-of-the-art open-source MoE LLMs and show that our adaptive expert silencing increases average attack success from 7.3% to 70.4%, reaching up to 86.3%, outperforming prior training-free MoE jailbreak methods. Moreover, bypassing guardrails typically requires silencing fewer than 20% of layer-wise experts while largely preserving general language utility. These results reveal a fundamental tension between efficiency-driven MoE design and robust safety alignment and motivate distributing safety mechanisms more robustly in future MoE LLMs with architecture- and routing-aware methods.

cs.CR↗

arXiv2026

IDEM Enough? Evolving Highly Nonlinear Idempotent Boolean Functions

Idempotent Boolean functions form a highly structured subclass of Boolean functions that is closely related to rotation symmetry under a normal-basis representation and to invariance under a fixed linear map in a polynomial basis. These functions are attractive as candidates for cryptographic design, yet their additional algebraic constraints make the search for high nonlinearity substantially more difficult than in the unconstrained case. In this work, we investigate evolutionary methods for constructing highly nonlinear idempotent Boolean functions for dimensions $n=5$ up to $n=12$ using a polynomial basis representation with canonical primitive polynomials. Our results show that the problem of evolving idempotent functions is difficult due to the disruptive nature of crossover and mutation operators. Next, we show that idempotence can be enforced by encoding the truth table on orbits, yielding a compact genome of size equal to the number of distinct squaring orbits.

cs.CR↗

arXiv2026

NegaBent, No Regrets: Evolving Spectrally Flat Boolean Functions

Negabent Boolean functions are defined by having a flat magnitude spectrum under the nega-Hadamard transform. They exist in both even and odd dimensions, and the subclass of functions that are simultaneously bent and negabent (bent-negabent) has attracted interest due to the combined optimal periodic and negaperiodic spectral properties. In this work, we investigate how evolutionary algorithms can be used to evolve (bent-)negabent Boolean functions. Our experimental results indicate that evolutionary algorithms, especially genetic programming, are a suitable approach for evolving negabent Boolean functions, and we successfully evolve such functions in all dimensions we consider.

cs.NE↗

arXiv2025

GateBreaker: Gate-Guided Attacks on Mixture-of-Expert LLMs

Mixture-of-Experts (MoE) architectures have advanced the scaling of Large Language Models (LLMs) by activating only a sparse subset of parameters per input, enabling state-of-the-art performance with reduced computational cost. As these models are increasingly deployed in critical domains, understanding and strengthening their alignment mechanisms is essential to prevent harmful outputs. However, existing LLM safety research has focused almost exclusively on dense architectures, leaving the unique safety properties of MoEs largely unexamined. The modular, sparsely-activated design of MoEs suggests that safety mechanisms may operate differently than in dense models, raising questions about their robustness. In this paper, we present GateBreaker, the first training-free, lightweight, and architecture-agnostic attack framework that compromises the safety alignment of modern MoE LLMs at inference time. GateBreaker operates in three stages: (i) gate-level profiling, which identifies safety experts disproportionately routed on harmful inputs, (ii) expert-level localization, which localizes the safety structure within safety experts, and (iii) targeted safety removal, which disables the identified safety structure to compromise the safety alignment. Our study shows that MoE safety concentrates within a small subset of neurons coordinated by sparse routing. Selective disabling of these neurons, approximately 3% of neurons in the targeted expert layers, significantly increases the averaged attack success rate (ASR) from 7.4% to 64.9% against the eight latest aligned MoE LLMs with limited utility degradation. These safety neurons transfer across models within the same family, raising ASR from 17.9% to 67.7% with one-shot transfer attack. Furthermore, GateBreaker generalizes to five MoE vision language models (VLMs) with 60.9% ASR on unsafe image inputs.

cs.CR↗

arXiv2025

MIMIR: Masked Image Modeling for Mutual Information-based Adversarial Robustness

Vision Transformers (ViTs) have emerged as a fundamental architecture and serve as the backbone of modern vision-language models. Despite their impressive performance, ViTs exhibit notable vulnerability to evasion attacks, necessitating the development of specialized Adversarial Training (AT) strategies tailored to their unique architecture. While a direct solution might involve applying existing AT methods to ViTs, our analysis reveals significant incompatibilities, particularly with state-of-the-art (SOTA) approaches such as Generalist (CVPR 2023) and DBAT (USENIX Security 2024). This paper presents a systematic investigation of adversarial robustness in ViTs and provides a novel theoretical Mutual Information (MI) analysis in its autoencoder-based self-supervised pre-training. Specifically, we show that MI between the adversarial example and its latent representation in ViT-based autoencoders should be constrained via derived MI bounds. Building on this insight, we propose a self-supervised AT method, MIMIR, that employs an MI penalty to facilitate adversarial pre-training by masked image modeling with autoencoders. Extensive experiments on CIFAR-10, Tiny-ImageNet, and ImageNet-1K show that MIMIR can consistently provide improved natural and robust accuracy, where MIMIR outperforms SOTA AT results on ImageNet-1K. Notably, MIMIR demonstrates superior robustness against unforeseen attacks and common corruption data and can also withstand adaptive attacks where the adversary possesses full knowledge of the defense mechanism. Our code and trained models are publicly available at: https://github.com/xiaoyunxxy/MIMIR.

cs.CV↗

arXiv2025

Towards Backdoor Stealthiness in Model Parameter Space

Recent research on backdoor stealthiness focuses mainly on indistinguishable triggers in input space and inseparable backdoor representations in feature space, aiming to circumvent backdoor defenses that examine these respective spaces. However, existing backdoor attacks are typically designed to resist a specific type of backdoor defense without considering the diverse range of defense mechanisms. Based on this observation, we pose a natural question: Are current backdoor attacks truly a real-world threat when facing diverse practical defenses? To answer this question, we examine 12 common backdoor attacks that focus on input-space or feature-space stealthiness and 17 diverse representative defenses. Surprisingly, we reveal a critical blind spot: Backdoor attacks designed to be stealthy in input and feature spaces can be mitigated by examining backdoored models in parameter space. To investigate the underlying causes behind this common vulnerability, we study the characteristics of backdoor attacks in the parameter space. Notably, we find that input- and feature-space attacks introduce prominent backdoor-related neurons in parameter space, which are not thoroughly considered by current backdoor attacks. Taking comprehensive stealthiness into account, we propose a novel supply-chain attack called Grond. Grond limits the parameter changes by a simple yet effective module, Adversarial Backdoor Injection (ABI), which adaptively increases the parameter-space stealthiness during the backdoor injection. Extensive experiments demonstrate that Grond outperforms all 12 backdoor attacks against state-of-the-art (including adaptive) defenses on CIFAR-10, GTSRB, and a subset of ImageNet. In addition, we show that ABI consistently improves the effectiveness of common backdoor attacks.

cs.CR↗

arXiv2025

Interpreting Emergent Features in Deep Learning-based Side-channel Analysis

Side-channel analysis (SCA) poses a real-world threat by exploiting unintentional physical signals to extract secret information from secure devices. Evaluation labs also use the same techniques to certify device security. In recent years, deep learning has emerged as a prominent method for SCA, achieving state-of-the-art attack performance at the cost of interpretability. Understanding how neural networks extract secrets is crucial for security evaluators aiming to defend against such attacks, as only by understanding the attack can one propose better countermeasures. In this work, we apply mechanistic interpretability to neural networks trained for SCA, revealing \textit{how} models exploit \textit{what} leakage in side-channel traces. We focus on sudden jumps in performance to reverse engineer learned representations, ultimately recovering secret masks and moving the evaluation process from black-box to white-box. Our results show that mechanistic interpretability can scale to realistic SCA settings, even when relevant inputs are sparse, model accuracies are low, and side-channel protections prevent standard input interventions.

cs.CR↗

arXiv2025

SoK: The Last Line of Defense: On Backdoor Defense Evaluation

Backdoor attacks pose a significant threat to deep learning models by implanting hidden vulnerabilities that can be activated by malicious inputs. While numerous defenses have been proposed to mitigate these attacks, the heterogeneous landscape of evaluation methodologies hinders fair comparison between defenses. This work presents a systematic (meta-)analysis of backdoor defenses through a comprehensive literature review and empirical evaluation. We analyzed 183 backdoor defense papers published between 2018 and 2025 across major AI and security venues, examining the properties and evaluation methodologies of these defenses. Our analysis reveals significant inconsistencies in experimental setups, evaluation metrics, and threat model assumptions in the literature. Through extensive experiments involving three datasets (MNIST, CIFAR-100, ImageNet-1K), four model architectures (ResNet-18, VGG-19, ViT-B/16, DenseNet-121), 16 representative defenses, and five commonly used attacks, totaling over 3\,000 experiments, we demonstrate that defense effectiveness varies substantially across different evaluation setups. We identify critical gaps in current evaluation practices, including insufficient reporting of computational overhead and behavior under benign conditions, bias in hyperparameter selection, and incomplete experimentation. Based on our findings, we provide concrete challenges and well-motivated recommendations to standardize and improve future defense evaluations. Our work aims to equip researchers and industry practitioners with actionable insights for developing, assessing, and deploying defenses to different systems.

cs.CR↗

arXiv2025

On Counts and Densities of Homogeneous Bent Functions: An Evolutionary Approach

Boolean functions with strong cryptographic properties, such as high nonlinearity and algebraic degree, are important for the security of stream and block ciphers. These functions can be designed using algebraic constructions or metaheuristics. This paper examines the use of Evolutionary Algorithms (EAs) to evolve homogeneous bent Boolean functions, that is, functions whose algebraic normal form contains only monomials of the same degree and that are maximally nonlinear. We introduce the notion of density of homogeneous bent functions, facilitating the algorithmic design that results in finding quadratic and cubic bent functions in different numbers of variables.

cs.NE↗

arXiv2025

CatBack: Universal Backdoor Attacks on Tabular Data via Categorical Encoding

Backdoor attacks in machine learning have drawn significant attention for their potential to compromise models stealthily, yet most research has focused on homogeneous data such as images. In this work, we propose a novel backdoor attack on tabular data, which is particularly challenging due to the presence of both numerical and categorical features. Our key idea is a novel technique to convert categorical values into floating-point representations. This approach preserves enough information to maintain clean-model accuracy compared to traditional methods like one-hot or ordinal encoding. By doing this, we create a gradient-based universal perturbation that applies to all features, including categorical ones. We evaluate our method on five datasets and four popular models. Our results show up to a 100% attack success rate in both white-box and black-box settings (including real-world applications like Vertex AI), revealing a severe vulnerability for tabular data. Our method is shown to surpass the previous works like Tabdoor in terms of performance, while remaining stealthy against state-of-the-art defense mechanisms. We evaluate our attack against Spectral Signatures, Neural Cleanse, Beatrix, and Fine-Pruning, all of which fail to defend successfully against it. We also verify that our attack successfully bypasses popular outlier detection mechanisms.

cs.LG↗

arXiv2025

NoMod: A Non-modular Attack on Module Learning With Errors

The advent of quantum computing threatens classical public-key cryptography, motivating NIST's adoption of post-quantum schemes such as those based on the Module Learning With Errors (Module-LWE) problem. We present NoMod ML-Attack, a hybrid white-box cryptanalytic method that circumvents the challenge of modeling modular reduction by treating wrap-arounds as statistical corruption and casting secret recovery as robust linear estimation. Our approach combines optimized lattice preprocessing--including reduced-vector saving and algebraic amplification--with robust estimators trained via Tukey's Biweight loss. Experiments show NoMod achieves full recovery of binary secrets for dimension $n = 350$, recovery of sparse binomial secrets for $n = 256$, and successful recovery of sparse secrets in CRYSTALS-Kyber settings with parameters $(n, k) = (128, 3)$ and $(256, 2)$. We release our implementation in an anonymous repository https://anonymous.4open.science/r/NoMod-3BD4.

cs.CR↗

arXiv2025

The SkipSponge Attack: Sponge Weight Poisoning of Deep Neural Networks

Sponge attacks aim to increase the energy consumption and computation time of neural networks. In this work, we present a novel sponge attack called SkipSponge. SkipSponge is the first sponge attack that is performed directly on the parameters of a pretrained model using only a few data samples. Our experiments show that SkipSponge can successfully increase the energy consumption of image classification models, GANs, and autoencoders, requiring fewer samples than the state-of-the-art sponge attacks (Sponge Poisoning). We show that poisoning defenses are ineffective if not adjusted specifically for the defense against SkipSponge (i.e., they decrease target layer bias values) and that SkipSponge is more effective on the GANs and the autoencoders than Sponge Poisoning. Additionally, SkipSponge is stealthy as it does not require significant changes to the victim model's parameters. Our experiments indicate that SkipSponge can be performed even when an attacker has access to less than 1% of the entire training dataset and reaches up to 13% energy increase.

cs.CR↗

arXiv2025

Backdoor Attacks on Transformers for Tabular Data: An Empirical Study

Deep Neural Networks (DNNs) have shown great promise in various domains. However, vulnerabilities associated with DNN training, such as backdoor attacks, are a significant concern. These attacks involve the subtle insertion of triggers during model training, allowing for manipulated predictions. More recently, DNNs used with tabular data have gained increasing attention due to the rise of transformer models. Our research presents a comprehensive analysis of backdoor attacks on tabular data using DNNs, mainly focusing on transformers. We propose a novel approach for trigger construction: in-bounds attack, which provides excellent attack performance while maintaining stealthiness. Through systematic experimentation across benchmark datasets, we uncover that transformer-based DNNs for tabular data are highly susceptible to backdoor attacks, even with minimal feature value alterations. We also verify that these attacks can be generalized to other models, like XGBoost and DeepFM. Our results demonstrate up to 100% attack success rate with negligible clean accuracy drop. Furthermore, we evaluate several defenses against these attacks, identifying Spectral Signatures as the most effective. Still, our findings highlight the need to develop tabular data-specific countermeasures to defend against backdoor attacks.

cs.CR↗

arXiv2025

A Systematic Study on the Design of Odd-Sized Highly Nonlinear Boolean Functions via Evolutionary Algorithms

This paper focuses on the problem of evolving Boolean functions of odd sizes with high nonlinearity, a property of cryptographic relevance. Despite its simple formulation, this problem turns out to be remarkably difficult. We perform a systematic evaluation by considering three solution encodings and four problem instances, analyzing how well different types of evolutionary algorithms behave in finding a maximally nonlinear Boolean function. Our results show that genetic programming generally outperforms other evolutionary algorithms, although it falls short of the best-known results achieved by ad-hoc heuristics. Interestingly, by adding local search and restricting the space to rotation symmetric Boolean functions, we show that a genetic algorithm with the bitstring encoding manages to evolve a $9$-variable Boolean function with nonlinearity 241.

cs.NE↗