arXiv preprints from January 1, 2026 through September 26, 2026 — 12:53:54 EST
Multimodal Large Language Models (MLLMs) have become widely deployed, yet their safety alignment remains fragile under adversarial inputs. Previous work has shown that increasing inference steps can disrupt safety mechanisms and lead MLLMs to generate attacker-desired harmful content. However, most existing attacks focus on increasing...
We investigate superradiant scattering off a rotating regular black hole described by a black-bounce metric which generalizes the Kerr spacetime of mass $M$ and specific angular momentum $a$ through a regularization parameter $p$ and two deformation exponents $(k,n)$. Focusing on massless $(\ell,m)=(1,1)$ scalar modes, we explore the...
Let p be any prime, and $p^(ν_p(n!))$ the maximal power of $p$ dividing $n!$. It is proved that there exists a positive integer $n_0$, which depends only on $p$, such that $q^(ν_q(n!)) < p^(ν_p(n!))$ for all $n \ge n_0$ and all primes $q > p$. For twin primes $p$ and $q = p + 2$ it is proved that the minimal $n_0$ satisfying...
This study highlights the potential of image-based reinforcement learning methods for addressing swarm-related tasks. In multi-agent reinforcement learning, effective policy learning depends on how agents sense, interpret, and process inputs. Traditional approaches often rely on handcrafted feature extraction or raw vector-based...
We study the stable translation length of homeomorphisms of a surface acting on the fine nonseparating curve graph and compare it to the stable translation lengths of its finite approximations - mapping classes relative to a finite invariant set - acting on the nonseparating curve graph. We prove that the stable translation length of...
We show that Activated Random Walk on $\mathbb{Z}$ is explosive above criticality. That is, activating a single particle in a supercritical state of sleeping particles triggers an infinite avalanche of activity with positive probability. This extends the same result recently proven by Brown, Hoffman, and Son for i.i.d. initial...
Molecular subtyping of PDAC into basal-like and classical has established prognostic and predictive value. However, its use in clinical practice is limited by cost, turnaround time, and tissue requirements, thereby restricting its application in the management of PDAC. We introduce PanSubNet, an interpretable deep learning framework...
We present a tabletop demonstration of dynamic stabilization and ponderomotive-like trapping using a pair of sinusoidally-driven anti-Helmholtz coils and a suspended permanent magnet. The oscillating field produces a rapid micromotion superimposed on a slower secular oscillation, with micromotion amplitude increasing with displacement...
With the arrival of JWST observations of the TRAPPIST-1 planets, it is timely to reassess the contribution of tidal heating to their heat budget. JWST thermal phase curves could reveal endogenic heating through an anomalously high nightside temperature, providing an opportunity to estimate tidal heating. In this study, we revisit the...
We introduce an implementation of a non-degenerate parametric amplifier in which the signal and idler modes, respectively, a microwave mode and an ensemble of spins (e.g., nitrogen-vacancy centers in diamond), are operated in their linear regime. This paramp, which amplifies signals in both parts at room and cryogenic temperatures,...
Given a projective variety $X$ and a very ample line bundle $\mathcal{L}$ on $X$, we classify for which $X$ and $\mathcal{L}$ the twisted syzygies and twisted dual syzygies bundles are Ulrich with respect to the polarizations $\mathcal{L}^a$. We obtain some partial results when considering an arbitrary polarization $H$.
This paper presents a method to identify electrons using the Cherenkov light emitted when a charged particle travels in air and photons are detected with a Silicon PhotoMultiplier (SiPM). The analysis is based on a photon-counting approach using SPAD cells and uses data collected during a test beam at CERN PS. The results are well...
A complex potential is a holomorphic function $Ω:\mathbb{C} \to \mathbb{C}$ whose real and imaginary parts generate a pair of orthogonal foliations, representing the equipotential lines and the streamlines of $\dot{z} = \overline{Ω'(z)}$. In this work, we generalize the concept of potential to the broader class of dynamical systems of...
We present a systematic formalization of Tigrinya cardinal and ordinal number verbalization, addressing a gap in computational resources for the language. This work documents the canonical rules governing the expression of numerical values in spoken Tigrinya, including the conjunction system, scale words, and special cases for dates,...
We construct a class of homogeneous Cantor-Moran measures with all contraction ratios being reciprocal of integers, and prove that they are pointwise absolutely normal. Our approach relies on methods developed by Davenport, Erd{ő}s, and LeVeque \cite{DEL1963} and properties of the order of integers in the multiplicative groups. The...
Artificial intelligence (AI) systems increasingly achieve expert-level predictive accuracy in healthcare, yet improvements in model performance often fail to produce corresponding gains in patient outcomes. We term this disconnect the allocation gap and provide a decision-theoretic explanation by modelling healthcare delivery as a...
Large language models (LLMs) are increasingly trained on massive, heterogeneous text corpora, raising serious concerns about the unauthorised use of proprietary or personal data during model training. In this work, we address the problem of data protection against unwanted model learning in a realistic black-box setting. We propose...
Vision-Language Models (VLMs) have achieved strong performance on standard vision-language benchmarks, yet often rely on surface-level recognition rather than deeper reasoning. We propose visual word puzzles as a challenging alternative, as they require discovering implicit visual cues, generating and revising hypotheses, and mapping...
We present a controlled study of multi-hop contextual reasoning in large language models, providing a clean demonstration of the task-method dissociation: rule-based pattern matching achieves 100% success on structured information retrieval but only 6.7% on tasks requiring cross-document reasoning, while LLM-based multi-agent systems...
A multifractal model is used to analyze neutron evolution within a reactor. For chain reactions, various characteristics of multifractal neutron behavior have been determined. These include the dimension of the multifractal carrier, information and correlation dimensions, the entropy of the fractal set, maximum and minimum dimension...
In this work, we introduce and formalize the Zero-Knowledge Task Planning (ZKTP) problem, i.e., formulating a sequence of actions to achieve some goal without task-specific knowledge. Additionally, we present a first investigation and approach for ZKTP that leverages a large language model (LLM) to decompose natural language...
We present PIVONet (Physically-Informed Variational ODE Neural Network), a unified framework that integrates Neural Ordinary Differential Equations (Neuro-ODEs) with Continuous Normalizing Flows (CNFs) for stochastic fluid simulation and visualization. First, we demonstrate that a physically informed model, parameterized by CNF...
Procedural content generation has enabled vast virtual worlds through levels, maps, and quests, but large-scale character generation remains underexplored. We identify two alignment-induced biases in existing methods: a positive moral bias, where characters uniformly adopt agreeable stances (e.g. always saying lying is bad), and a...
We examine multiphoton interference with a symmetric $SU(N)$ beam splitter $S_N$, an extension of features of the $SU(2)$ 50/50 beam splitter extended Hong-Ou-Mandel (eHOM) effect, whereby one obtains a zero amplitude (probability) for the output coincidence state (defined by equal number of photons $n/N$ in each output port), when a...
The computational design of heteroatom-doped organic dyes for dye-sensitized solar cells (DSSCs) remains challenging, as predictive methods must accurately describe long-range charge-transfer (CT) excitations while remaining computationally efficient for systematic materials screening. In this work, we investigate the electronic...