Jailbreak transferability is the surprising phenomenon when an adversarial attack compromising one model also elicits harmful responses from other models. Despite widespread demonstrations, there is little consensus on why transfer is possible: is it a quirk of safety training, an artifact of model families, or a more fundamental property of representation learning? We present evidence that transferability emerges from shared representations rather than incidental flaws. Across 20 open-weight models and 33 jailbreak attacks, we find two factors that systematically shape transfer: (1) representational similarity under benign prompts, and (2) the strength of the jailbreak on the source model. To move beyond correlation, we show that deliberately increasing similarity through benign only distillation causally increases transfer. Our qualitative analyses reveal systematic transferability patterns across different types of jailbreaks. For example, persona-style jailbreaks transfer far more often than cipher-based prompts, consistent with the idea that natural-language attacks exploit models' shared representation space, whereas cipher-based attacks rely on idiosyncratic quirks that do not generalize. Together, these results reframe jailbreak transfer as a consequence of representation alignment rather than a fragile byproduct of safety training.

The developments over the last five decades concerning numerical discretisations of the incompressible Navier--Stokes equations have lead to reliable tools for their approximation: those include stable methods to properly address the incompressibility constraint, stable discretisations to account for convection dominated problems, efficient time (splitting) methods, and methods to tackle their nonlinear character. While these tools may successfully be applied to reliably simulate even more complex fluid flow PDE models, their understanding requires a fundamental revision in the case of stochastic fluid models, which are gaining increased importance nowadays. This work motivates and surveys optimally convergent numerical methods for the stochastic Stokes and Navier--Stokes equations that were obtained in the last decades. Furtheremore, we computationally illustrate the failure of some of those methods from the deterministic setting, if they are straight-forwardly applied to the stochastic case. In fact, we explain why some of these deterministic methods perform sub-optimally by highlighting crucial analytical differences between the deterministic and stochastic equations -- and how modifications of the deterministic methods restore their optimal performance if they properly address the probabilistic nature of the stochastic problem. Next to the numerical analysis of schemes, we propose a general benchmark of prototypic fluid flow problems driven by different types of noise to also compare new algorithms by simulations in terms of complexities, efficiencies, and possible limitations. The driving motivation is to reach a better comparison of simulations for new schemes in terms of accuracy and complexities, and to also complement theoretical performance studies for restricted settings of data by more realistic ones.

Explanation is necessary for humans to understand and accept decisions made by an AI system when the system's goal is known. It is even more important when the AI system makes decisions in multi-agent environments where the human does not know the systems' goals since they may depend on other agents' preferences. In such situations, explanations should aim to increase user satisfaction, taking into account the system's decision, the user's and the other agents' preferences, the environment settings and properties such as fairness, envy and privacy. Generating explanations that will increase user satisfaction is very challenging; to this end, we propose a new research direction: xMASE. We then review the state of the art and discuss research directions towards efficient methodologies and algorithms for generating explanations that will increase users' satisfaction from AI system's decisions in multi-agent environments.

Topic modeling seems to be almost synonymous with generating lists of top words to represent topics within large text corpora. However, deducing a topic from such list of individual terms can require substantial expertise and experience, making topic modelling less accessible to people unfamiliar with the particularities and pitfalls of top-word interpretation. A topic representation limited to top-words might further fall short of offering a comprehensive and easily accessible characterization of the various aspects, facets and nuances a topic might have. To address these challenges, we introduce GPTopic, a software package that leverages Large Language Models (LLMs) to create dynamic, interactive topic representations. GPTopic provides an intuitive chat interface for users to explore, analyze, and refine topics interactively, making topic modeling more accessible and comprehensive. The corresponding code is available here: this https URL.

Osteoarthritis (OA) is a multifaceted joint disease which poses significant socioeconomic burdens and remains a significant clinical challenge. Evidence suggests that structural and mechanical changes in subchondral bone influence the pathogenesis and development of OA, leading to diminished bone quality and cartilage degeneration. While changes in microstructure and tissue scale elastic properties are well reported, the tissue yield response of subchondral bone in OA and their correlation with compositional changes have not been investigated. Here, we performed quasistatic micropillar compression and nanoindentation within the subchondral bone plate and trabeculae of hydrated non-diseased (ND) and OA affected specimens retrieved from the distal tibia in vivo. The micropillars, extracted by laser ablation, exhibited a taper angle which mandated the use of an in silico micropillar compression routine to back-calculate elastic modulus and strength of the bone tissue that comprised each micropillar. Elastic modulus remained unchanged between ND and OA subchondral bone, whereas strength increased from 46.0 MPa to 57.3 MPa in OA subchondral trabecular bone but not in the bone plate. Micropillar matched Raman spectroscopy and quantitative backscattered electron imaging revealed mineralisation is the underlying determinant of elastic modulus and strength at the microscale. By combining micromechanical and tissue compositional analyses, we investigated how the mechanical properties are related and how these properties are affected in subchondral bone by OA. Our results may be of value in the development and optimisation of interventions used to alleviate the socioeconomic burdens associated with this debilitating joint disease.

Deep neural networks (DNNs) have proven to be highly effective in a variety of tasks, making them the go-to method for problems requiring high-level predictive power. Despite this success, the inner workings of DNNs are often not transparent, making them difficult to interpret or understand. This lack of interpretability has led to increased research on inherently interpretable neural networks in recent years. Models such as Neural Additive Models (NAMs) achieve visual interpretability through the combination of classical statistical methods with DNNs. However, these approaches only concentrate on mean response predictions, leaving out other properties of the response distribution of the underlying data. We propose Neural Additive Models for Location Scale and Shape (NAMLSS), a modelling framework that combines the predictive power of classical deep learning models with the inherent advantages of distributional regression while maintaining the interpretability of additive models. The code is available at the following link: this https URL

We show that the proportional clustering problem using the Droop quota for $k = 1$is equivalent to the$\beta$-plurality problem. We also show that the Plurality Veto rule can be used to select ($\sqrt{5} - 2$)-plurality points using only ordinal information about the metric space and resolve an open question of Kalayci et al. (AAAI 2024) by proving that $(2+\sqrt{5})$-proportionally fair clusterings can be found using purely ordinal information.

We study the election of sequences of committees, where in each of $\tau$ levels (e.g. modeling points in time) a committee consisting of $k$ candidates from a common set of $m$ candidates is selected. For each level, each of $n$ agents (voters) may nominate one candidate whose selection would satisfy her. We are interested in committees which are good with respect to the satisfaction per day and per agent. More precisely, we look for egalitarian or equitable committee sequences. While both guarantee that at least $x$ agents per day are satisfied, egalitarian committee sequences ensure that each agent is satisfied in at least $y$ levels while equitable committee sequences ensure that each agent is satisfied in exactly $y$ levels. We analyze the parameterized complexity of finding such committees for the parameters $n,m,k,\tau,x$, and $y$, as well as combinations thereof.

Transformers are the current state-of-the-art of natural language processing in many domains and are using traction within software engineering research as well. Such models are pre-trained on large amounts of data, usually from the general domain. However, we only have a limited understanding regarding the validity of transformers within the software engineering domain, i.e., how good such models are at understanding words and sentences within a software engineering context and how this improves the state-of-the-art. Within this article, we shed light on this complex, but crucial issue. We compare BERT transformer models trained with software engineering data with transformers based on general domain data in multiple dimensions: their vocabulary, their ability to understand which words are missing, and their performance in classification tasks. Our results show that for tasks that require understanding of the software engineering context, pre-training with software engineering data is valuable, while general domain models are sufficient for general language understanding, also within the software engineering domain.

WebAssembly has attracted great attention as a portable compilation target for programming languages. To facilitate in-depth studies about this technology, we have deployed Wasmizer, a tool that regularly mines GitHub projects and makes an up-to-date dataset of WebAssembly sources and their binaries publicly available. Presently, we have collected 2 540 C and C++ projects that are highly-related to WebAssembly, and built a dataset of 8 915 binaries that are linked to their source projects. To demonstrate an application of this dataset, we have investigated the presence of eight WebAssembly compilation smells in the wild.

This paper describes our work on demonstrating verification technologies on a flight-critical system of realistic functionality, size, and complexity. Our work targeted a commercial aircraft control system named Transport Class Model (TCM), and involved several stages: formalizing and disambiguating requirements in collaboration with do- main experts; processing models for their use by formal verification tools; applying compositional techniques at the architectural and component level to scale verification. Performed in the context of a major NASA milestone, this study of formal verification in practice is one of the most challenging that our group has performed, and it took several person months to complete it. This paper describes the methodology that we followed and the lessons that we learned.

Judgment aggregation is a framework to aggregate individual opinions on multiple, logically connected issues into a collective outcome. These opinions are cast by judges, which can be for example referees, experts, advisors or jurors, depending on the application and context. It is open to manipulative attacks such as \textsc{Manipulation} where judges cast their judgments strategically. Previous works have shown that most computational problems corresponding to these manipulative attacks are \NP-hard. This desired computational barrier, however, often relies on formulas that are either of unbounded size or of complex structure. We revisit the computational complexity for various \textsc{Manipulation} and \textsc{Bribery} problems in premise-based judgment aggregation, now focusing on simple and realistic formulas. We restrict all formulas to be clauses that are (positive) monotone, Horn-clauses, or have bounded length. For basic variants of \textsc{Manipulation}, we show that these restrictions make several variants, which were in general known to be \NP-hard, polynomial-time solvable. Moreover, we provide a P vs.\ NP dichotomy for a large class of clause restrictions (generalizing monotone and Horn clauses) by showing a close relationship between variants of \textsc{Manipulation} and variants of \textsc{Satisfiability}. For Hamming distance based \textsc{Manipulation}, we show that \NP-hardness even holds for positive monotone clauses of length three, but the problem becomes polynomial-time solvable for positive monotone clauses of length two. For \textsc{Bribery}, we show that \NP-hardness even holds for positive monotone clauses of length two, but it becomes polynomial-time solvable for the same clause set if there is a constant budget.

We propose a two-point flux approximation finite-volume scheme for a stochastic non-linear parabolic equation with a multiplicative noise. The time discretization is implicit except for the stochastic noise term in order to be compatible with stochastic integration in the sense of Itô. We show existence and uniqueness of solutions to the scheme and the appropriate measurability for stochastic integration follows from the uniqueness of approximate solutions.

To aggregate rankings into a social ranking, one can use scoring systems such as Plurality, Veto, and Borda. We distinguish three types of methods: ranking by score, ranking by repeatedly choosing a winner that we delete and rank at the top, and ranking by repeatedly choosing a loser that we delete and rank at the bottom. The latter method captures the frequently studied voting rules Single Transferable Vote (aka Instant Runoff Voting), Coombs, and Baldwin. In an experimental analysis, we show that the three types of methods produce different rankings in practice. We also provide evidence that sequentially selecting winners is most suitable to detect the "true" ranking of candidates. For different rules in our classes, we then study the (parameterized) computational complexity of deciding in which positions a given candidate can appear in the chosen ranking. As part of our analysis, we also consider the Winner Determination problem for STV, Coombs, and Baldwin and determine their complexity when there are few voters or candidates.

Extracting and identifying latent topics in large text corpora has gained increasing importance in Natural Language Processing (NLP). Most models, whether probabilistic models similar to Latent Dirichlet Allocation (LDA) or neural topic models, follow the same underlying approach of topic interpretability and topic extraction. We propose a method that incorporates a deeper understanding of both sentence and document themes, and goes beyond simply analyzing word frequencies in the data. This allows our model to detect latent topics that may include uncommon words or neologisms, as well as words not present in the documents themselves. Additionally, we propose several new evaluation metrics based on intruder words and similarity measures in the semantic space. We present correlation coefficients with human identification of intruder words and achieve near-human level results at the word-intrusion task. We demonstrate the competitive performance of our method with a large benchmark study, and achieve superior results compared to state-of-the-art topic modeling and document clustering models.

In recent years, stochastic effects have become increasingly relevant for describing fluid behaviour, particularly in the context of turbulence. The most important model for inviscid fluids in computational fluid dynamics are the Euler equations of gas dynamics which we focus on in this paper. To take stochastic effects into account, we incorporate a stochastic forcing term in the momentum equation of the Euler system. To solve the extended system, we apply an entropy dissipative discontinuous Galerkin spectral element method including the Finite Volume setting, adjust it to the stochastic Euler equations and analyze its convergence properties. Our analysis is grounded in the concept of dissipative martingale solutions, as recently introduced by Moyo (J. Diff. Equ. 365, 408-464, 2023). Assuming no vacuum formation and bounded total energy, we proof that our scheme converges in law to a dissipative martingale solution. During the lifespan of a pathwise strong solution, we achieve convergence of at least order 1/2, measured by the expected $L^1$ norm of the relative energy. The results built a counterpart of corresponding results in the deterministic case. In numerical simulations, we show the robustness of our scheme, visualise different stochastic realizations and analyze our theoretical findings.

There is a pressing demand for robust, high-order baseline schemes for conservation laws that minimize reliance on supplementary stabilization. In this work, we respond to this demand by developing new baseline schemes within a nodal discontinuous Galerkin (DG) framework, utilizing upwind summation-by-parts (USBP) operators and flux vector splittings. To this end, we demonstrate the existence of USBP operators on arbitrary grid points and provide a straightforward procedure for their construction. Our method encompasses a broader class of USBP operators, not limited to equidistant grid points. This approach facilitates the development of novel USBP operators on Legendre--Gauss--Lobatto (LGL) points, which are suited for nodal discontinuous Galerkin (DG) methods. The resulting DG-USBP operators combine the strengths of traditional summation-by-parts (SBP) schemes with the benefits of upwind discretizations, including inherent dissipation mechanisms. Through numerical experiments, ranging from one-dimensional convergence tests to multi-dimensional curvilinear and under-resolved flow simulations, we find that DG-USBP operators, when integrated with flux vector splitting methods, foster more robust baseline schemes without excessive artificial dissipation.

We inspected 45 actively deployed Operational Technology (OT) product families from ten major vendors and found that every system suffers from at least one trivial vulnerability. We reported a total of 53 weaknesses, stemming from insecure by design practices or basic security design failures. They enable attackers to take a device offline, manipulate its operational parameters, and execute arbitrary code without any constraint. We discuss why vulnerable products are often security certified and appear to be more secure than they actually are, and we explain complicating factors of OT risk management.

Inhalers spray over 100 million drug particles into the mouth, where a significant portion of the drug may deposit. Understanding how the complex interplay between particle and solid phases influence deposition is crucial for optimising treatments. Existing modelling studies neglect any effect of particle momentum on the fluid (one-way coupling), which may cause poor prediction of forces acting on particles. In this study, we simulate a realistic number of particles (up to 160 million) in a patient-specific geometry. We study the effect of momentum transfer from particles to the fluid (two-way coupling) and particle-particle interactions (four-way coupling) on deposition. We also explore the effect of tracking groups of particles (`parcels') to lower computational cost. Upper airway deposition fraction increased from 0.33 (one-way coupled) to 0.87 with two-way coupling and $10 \mu m$ particle diameter. Four-way coupling lowers upper airway deposition by approximately 10% at $100 \mu g$ dosages. We use parcel modelling to study deposition of $4 - 20 \mu m$ particles, observing significant influence of two-way coupling in each simulation. These results show that future studies should model realistic dosages for accurate prediction of deposition which may inform clinical decision-making.