Tootfinder

SafeCOMM: What about Safety Alignment in Fine-Tuned Telecom Large Language Models?
Aladin Djuhera, Swanand Ravindra Kadhe, Farhan Ahmed, Syed Zawad, Holger Boche, Walid Saad
https://arxiv.org/abs/2506.00062

SafeCOMM: What about Safety Alignment in Fine-Tuned Telecom Large Language Models?
Fine-tuning large language models (LLMs) for telecom tasks and datasets is a common practice to adapt general-purpose models to the telecom domain. However, little attention has been paid to how this process may compromise model safety. Recent research has shown that even benign fine-tuning can degrade the safety alignment of LLMs, causing them to respond to harmful or unethical user queries. In this paper, we investigate this issue for telecom-tuned LLMs using three representative datasets fea…

A busy week! some activities less futile than others...? stay tuned for updates
* last week of semester 🐘
* submitted revised paper to journal (and arXiv) after minor corrections, thanks to #Reviewer2
* submitted abstract for #ICM2026 short talk
* working with co-organizer to …

⛓️‍💥 Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations
#ai #llm

Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations
Modern tokenizers employ deterministic algorithms to map text into a single "canonical" token sequence, yet the same string can be encoded as many non-canonical tokenizations using the tokenizer vocabulary. In this work, we investigate the robustness of LMs to text encoded with non-canonical tokenizations entirely unseen during training. Surprisingly, when evaluated across 20 benchmarks, we find that instruction-tuned models retain up to 93.4% of their original performance when given a randomly…

Linearly Decoding Refused Knowledge in Aligned Language Models
Aryan Shrivastava, Ari Holtzman
https://arxiv.org/abs/2507.00239 https://

Linearly Decoding Refused Knowledge in Aligned Language Models
Most commonly used language models (LMs) are instruction-tuned and aligned using a combination of fine-tuning and reinforcement learning, causing them to refuse users requests deemed harmful by the model. However, jailbreak prompts can often bypass these refusal mechanisms and elicit harmful responses. In this work, we study the extent to which information accessed via jailbreak prompts is decodable using linear probes trained on LM hidden states. We show that a great deal of initially refused …

Understanding Overadaptation in Supervised Fine-Tuning: The Role of Ensemble Methods
Yifan Hao, Xingyuan Pan, Hanning Zhang, Chenlu Ye, Rui Pan, Tong Zhang
https://arxiv.org/abs/2506.01901

Understanding Overadaptation in Supervised Fine-Tuning: The Role of Ensemble Methods
Supervised fine-tuning (SFT) on domain-specific data is the dominant approach for adapting foundation models to specialized tasks. However, it has been observed that SFT models tend to forget knowledge acquired during pretraining. In vision models, ensembling a pretrained model with its fine-tuned counterpart has been shown to mitigate this issue. In this work, we demonstrate that the same holds for language models, and, more strikingly, we observe an overadaptation phenomenon: the ensemble mod…

Trojan Horse Hunt in Time Series Forecasting for Space Operations
Krzysztof Kotowski, Ramez Shendy, Jakub Nalepa, Przemys{\l}aw Biecek, Piotr Wilczy\'nski, Agata Kaczmarek, Dawid P{\l}udowski, Artur Janicki, Evridiki Ntagiou
https://arxiv.org/abs/2506.01849

Trojan Horse Hunt in Time Series Forecasting for Space Operations
This competition hosted on Kaggle (https://www.kaggle.com/competitions/trojan-horse-hunt-in-space) is the first part of a series of follow-up competitions and hackathons related to the "Assurance for Space Domain AI Applications" project funded by the European Space Agency (https://assurance-ai.space-codev.org/). The competition idea is based on one of the real-life AI security threats identified within the project -- the adversarial poisoning of continuously fine-tuned satellite telemetry fore…

This https://arxiv.org/abs/2411.16746 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCR_…

LoBAM: LoRA-Based Backdoor Attack on Model Merging
Model merging is an emerging technique that integrates multiple models fine-tuned on different tasks to create a versatile model that excels in multiple domains. This scheme, in the meantime, may open up backdoor attack opportunities where one single malicious model can jeopardize the integrity of the merged model. Existing works try to demonstrate the risk of such attacks by assuming substantial computational resources, focusing on cases where the attacker can fully fine-tune the pre-trained m…

Temperature and Magnetic-Field Dependence of Energy Relaxation in a Fluxonium Qubit
Lamia Ateshian, Max Hays, David A. Rower, Helin Zhang, Kate Azar, R\'eouven Assouly, Leon Ding, Michael Gingras, Hannah Stickler, Bethany M. Niedzielski, Mollie E. Schwartz, Terry P. Orlando, Joel \^I-j. Wang, Simon Gustavsson, Jeffrey A. Grover, Kyle Serniak, William D. Oliver

Temperature and Magnetic-Field Dependence of Energy Relaxation in a Fluxonium Qubit
Noise from material defects at device interfaces is known to limit the coherence of superconducting circuits, yet our understanding of the defect origins and noise mechanisms remains incomplete. Here we investigate the temperature and in-plane magnetic-field dependence of energy relaxation in a low-frequency fluxonium qubit, where the sensitivity to flux noise and charge noise arising from dielectric loss can be tuned by applied flux. We observe an approximately linear scaling of flux noise wit…

Evo-0: Vision-Language-Action Model with Implicit Spatial Understanding
Tao Lin, Gen Li, Yilei Zhong, Yanwen Zou, Bo Zhao
https://arxiv.org/abs/2507.00416 …

Evo-0: Vision-Language-Action Model with Implicit Spatial Understanding
Vision-Language-Action (VLA) models have emerged as a promising framework for enabling generalist robots capable of perceiving, reasoning, and acting in the real world. These models usually build upon pretrained Vision-Language Models (VLMs), which excel at semantic understanding due to large-scale text pretraining. However, VLMs typically lack precise spatial understanding capabilities, as they are primarily tuned on 2D image-text pairs without 3D supervision. To address this limitation, recen…

Real-Time Emergency Vehicle Siren Detection with Efficient CNNs on Embedded Hardware
Marco Giordano, Stefano Giacomelli, Claudia Rinaldi, Fabio Graziosi
https://arxiv.org/abs/2507.01563

Real-Time Emergency Vehicle Siren Detection with Efficient CNNs on Embedded Hardware
We present a full-stack emergency vehicle (EV) siren detection system designed for real-time deployment on embedded hardware. The proposed approach is based on E2PANNs, a fine-tuned convolutional neural network derived from EPANNs, and optimized for binary sound event detection under urban acoustic conditions. A key contribution is the creation of curated and semantically structured datasets - AudioSet-EV, AudioSet-EV Augmented, and Unified-EV - developed using a custom AudioSet-Tools framework…

GLoSS: Generative Language Models with Semantic Search for Sequential Recommendation
Krishna Acharya, Aleksandr V. Petrov, Juba Ziani
https://arxiv.org/abs/2506.01910

GLoSS: Generative Language Models with Semantic Search for Sequential Recommendation
We propose Generative Low-rank language model with Semantic Search (GLoSS), a generative recommendation framework that combines large language models with dense retrieval for sequential recommendation. Unlike prior methods such as GPT4Rec, which rely on lexical matching via BM25, GLoSS uses semantic search to retrieve relevant items beyond lexical matching. For query generation, we employ 4-bit quantized LlaMA-3 models fine-tuned with low-rank adaptation (LoRA), enabling efficient training and …

Save the date!
The fourth annual End of Life Doula International Research Group symposium is online September 5-6. Global south inclusion is a focus, with a diversity of international practice approaches included.
Registration opens soon, stay tuned!
#EOLD #EOLDIRG

This https://arxiv.org/abs/2406.14567 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csGR_…

DragPoser: Motion Reconstruction from Variable Sparse Tracking Signals via Latent Space Optimization
High-quality motion reconstruction that follows the user's movements can be achieved by high-end mocap systems with many sensors. However, obtaining such animation quality with fewer input devices is gaining popularity as it brings mocap closer to the general public. The main challenges include the loss of end-effector accuracy in learning-based approaches, or the lack of naturalness and smoothness in IK-based solutions. In addition, such systems are often finely tuned to a specific number of t…

Pitfalls of Evaluating Language Models with Open Benchmarks
Md. Najib Hasan (Wichita State University), Mohammad Fakhruddin Babar (Washington State University), Souvika Sarkar (Wichita State University), Monowar Hasan (Washington State University), Santu Karmaker (University of Central Florida)
https://arxiv.org/abs/2507.00460…

Pitfalls of Evaluating Language Models with Open Benchmarks
Open Large Language Model (LLM) benchmarks, such as HELM and BIG-bench, offer standardized, transparent protocols that facilitate the fair comparison, reproducibility, and iterative advancement of Language Models (LMs). However, their openness also introduces critical and underexplored pitfalls. This study exposes these weaknesses by systematically constructing ``cheating'' models -- smaller variants of BART, T5, and GPT-2 fine-tuned directly on public test sets -- which achieve top rankings on…

Versatile multi-q antiferromagnetic charge order in correlated vdW metals
Y. Fujisawa, P. Wu, R. Okuma, B. R. M. Smith, D. Ueta, R. Kobayashi, N. Maekawa, T. Nakamura, C-H. Hsu, Chandan De, N. Tomoda, T. Higashihara, K. Morishita, T. Kato, Z. Y. Wang, Y. Okada
https://arxiv.org/abs/2507.00750

Versatile multi-q antiferromagnetic charge order in correlated vdW metals
Following the discovery of graphene, interest in van der Waals (vdW) materials has surged; yet, advancing "beyond graphene" physics requires the development of quantum material platforms that host versatile many-body states. Using scanning tunneling microscopy and spectroscopy at 300 mK, we uncover two competing states in vdW metal CeTe3: charge-ordered in-plane antiferromagnetic phases forming stripe and checkerboard patterns. Remarkably, the competition between them is tuned through a modest …

LearnAFE: Circuit-Algorithm Co-design Framework for Learnable Audio Analog Front-End
Jinhai Hu, Zhongyi Zhang, Cong Sheng Leow, Wang Ling Goh, Yuan Gao
https://arxiv.org/abs/2507.00755

LearnAFE: Circuit-Algorithm Co-design Framework for Learnable Audio Analog Front-End
This paper presents a circuit-algorithm co-design framework for learnable analog front-end (AFE) in audio signal classification. Designing AFE and backend classifiers separately is a common practice but non-ideal, as shown in this paper. Instead, this paper proposes a joint optimization of the backend classifier with the AFE's transfer function to achieve system-level optimum. More specifically, the transfer function parameters of an analog bandpass filter (BPF) bank are tuned in a signal-to-no…

Solving inverse problems of Type IIB flux vacua with conditional generative models
Sven Krippendorf, Zhimei Liu
https://arxiv.org/abs/2506.22551 https://…

Solving inverse problems of Type IIB flux vacua with conditional generative models
We address the inverse problem in Type IIB flux compactifications of identifying flux vacua with targeted phenomenological properties such as specific superpotential values or tadpole constraints using conditional generative models. These machine learning techniques overcome computational bottlenecks in traditional approaches such as rejection sampling and Markov Chain Monte Carlo (MCMC), which struggle to generate rare, finely-tuned vacua. As a proof of concept, we demonstrate that conditional…

I've been pretty quiet here lately, but I've not been goofing off. I am leveraging what I've learned about security and privacy over the last two decades to create guides for the non-technical person. Easy to read (hopefully) guides on messaging, email, browsing and other topics that will make it easy for anyone to improve their Internet privacy. Stay tuned.
By the way, this is not a commercial effort. The guides will be free to download, use and share.
Why am I doin…

Selenization of V2O5/WO3 Bilayers for Tuned Optoelectronic Response of WSe2 Films
Abhishek Bajgain, Santu Prasad Jana, Alexander Samokhvalov, Thomas Parker, John Derek Demaree, Ramesh C. Budhani
https://arxiv.org/abs/2505.22800

Selenization of V2O5/WO3 Bilayers for Tuned Optoelectronic Response of WSe2 Films
Scalable and controlled doping of two-dimensional transition metal dichalcogenides is essential for tuning their electronic and optoelectronic properties. In this work, we demonstrate a robust approach for substitution of vanadium in tungsten diselenide (WSe$_2$) via the selenization of pre-deposited V$_2$O$_5$/WO$_3$ thin films. By adjusting the thickness of the vanadium oxide layer, the V concentration in W$_{1-x}$V$_x$Se$_2$ is systematically varied. Electrical measurements on field-effect t…

Reinforcement Fine-Tuned Large Language Models for Next POI Recommendation
Peibo Li, Shuang Ao, Hao Xue, Yang Song, Maarten de Rijke, Johan Barth\'elemy, Tomasz Bednarz, Flora D. Salim
https://arxiv.org/abs/2506.21599

Reinforcement Fine-Tuned Large Language Models for Next POI Recommendation
Large language models (LLMs) have been adopted for next point-of-interest (POI) recommendation tasks. Typical LLM-based recommenders fall into two categories: prompt-based and supervised fine-tuning (SFT)-based models. Prompt-based models generally offer greater output flexibility but deliver lower accuracy, whereas SFT-based models achieve higher performance yet face a fundamental mismatch: next POI recommendation data does not naturally suit supervised fine-tuning. In SFT, the model is traine…

ACTLLM: Action Consistency Tuned Large Language Model
Jing Bi, Lianggong Bruce Wen, Zhang Liu, Chenliang Xu
https://arxiv.org/abs/2506.21250 https://

ACTLLM: Action Consistency Tuned Large Language Model
This paper introduces ACTLLM (Action Consistency Tuned Large Language Model), a novel approach for robot manipulation in dynamic environments. Traditional vision-based systems often struggle to learn visual representations that excel in both task execution and spatial reasoning, thereby limiting their adaptability in dynamic environments. ACTLLM addresses these challenges by harnessing language to craft structured scene descriptors, providing a uniform interface for both spatial understanding a…

Today we're taking an (unplanned) break from streaming, but we'll be back next week with a new guest and a really fun topic!
Stay tuned

Fine-Tuned Supernova or Failed Explosion? Decoding the Origins of the G3425 Binary
Zsolt Regaly, Viktoria Frohlich, Jozsef Vinko
https://arxiv.org/abs/2506.20386

Fine-Tuned Supernova or Failed Explosion? Decoding the Origins of the G3425 Binary
A binary system (G3425) consisting of a massive unseen component and a red giant star on a nearly circular orbit was recently discovered. The formation of such a system is puzzling because orbital stability generally breaks down due to the large mass loss from the system caused by the SN explosion while forming the unseen component. Analytical solutions of the variable-mass two-body problem suggest that the explosion should have occurred when the component was close to its apocenter to explain …

ChemActor: Enhancing Automated Extraction of Chemical Synthesis Actions with LLM-Generated Data
Yu Zhang, Ruijie Yu, Jidong Tian, Feng Zhu, Jiapeng Liu, Xiaokang Yang, Yaohui Jin, Yanyan Xu
https://arxiv.org/abs/2506.23520

ChemActor: Enhancing Automated Extraction of Chemical Synthesis Actions with LLM-Generated Data
With the increasing interest in robotic synthesis in the context of organic chemistry, the automated extraction of chemical procedures from literature is critical. However, this task remains challenging due to the inherent ambiguity of chemical language and the high cost of human annotation required for developing reliable computer-aided extraction protocols. Here, we present ChemActor, a fully fine-tuned large language model (LLM), as a chemical executor to convert between unstructured experim…

Model-free Forecasting of Rogue Waves using Reservoir Computing
Abrari Noor Hasmi, Hadi Susanto
https://arxiv.org/abs/2506.21918 https://

Model-free Forecasting of Rogue Waves using Reservoir Computing
Recent research has demonstrated Reservoir Computing's capability to model various chaotic dynamical systems, yet its application to Hamiltonian systems remains relatively unexplored. This paper investigates the effectiveness of Reservoir Computing in capturing rogue wave dynamics from the nonlinear Schrödinger equation, a challenging Hamiltonian system with modulation instability. The model-free approach learns from breather simulations with five unstable modes. A properly tuned parallel Echo…

A Tool for Generating Exceptional Behavior Tests With Large Language Models
Linghan Zhong, Samuel Yuan, Jiyang Zhang, Yu Liu, Pengyu Nie, Junyi Jessy Li, Milos Gligoric
https://arxiv.org/abs/2505.22818

A Tool for Generating Exceptional Behavior Tests With Large Language Models
Exceptional behavior tests (EBTs) are crucial in software development for verifying that code correctly handles unwanted events and throws appropriate exceptions. However, prior research has shown that developers often prioritize testing "happy paths", e.g., paths without unwanted events over exceptional scenarios. We present exLong, a framework that automatically generates EBTs to address this gap. exLong leverages a large language model (LLM) fine-tuned from CodeLlama and incorporates reasoni…

A Detailed Factor Analysis for the Political Compass Test: Navigating Ideologies of Large Language Models
Sadia Kamal, Lalu Prasad Yadav Prakash, S M Rafiuddin, Mohammed Rakib, Arunkumar Bagavathi, Atriya Sen, Sagnik Ray Choudhury
https://arxiv.org/abs/2506.22493

A Detailed Factor Analysis for the Political Compass Test: Navigating Ideologies of Large Language Models
Political Compass Test (PCT) or similar questionnaires have been used to quantify LLM's political leanings. Building on a recent line of work that examines the validity of PCT tests, we demonstrate that variation in standard generation parameters does not significantly impact the models' PCT scores. However, external factors such as prompt variations and fine-tuning individually and in combination affect the same. Finally, we demonstrate that when models are fine-tuned on text datasets with hig…

Seq2Bind Webserver for Decoding Binding Hotspots directly from Sequences using Fine-Tuned Protein Language Models
Xiang Ma, Supantha Dey, Vaishnavey SR, Casey Zelinski, Qi Li, Ratul Chowdhury
https://arxiv.org/abs/2506.13830

Seq2Bind Webserver for Decoding Binding Hotspots directly from Sequences using Fine-Tuned Protein Language Models
Decoding protein-protein interactions (PPIs) at the residue level is crucial for understanding cellular mechanisms and developing targeted therapeutics. We present Seq2Bind Webserver, a computational framework that leverages fine-tuned protein language models (PLMs) to determine binding affinity between proteins and identify critical binding residues directly from sequences, eliminating the structural requirements that limit most affinity prediction tools. We fine-tuned four architectures inclu…

Self-correcting Reward Shaping via Language Models for Reinforcement Learning Agents in Games
Ant\'onio Afonso, Iolanda Leite, Alessandro Sestini, Florian Fuchs, Konrad Tollmar, Linus Gissl\'en
https://arxiv.org/abs/2506.23626

Self-correcting Reward Shaping via Language Models for Reinforcement Learning Agents in Games
Reinforcement Learning (RL) in games has gained significant momentum in recent years, enabling the creation of different agent behaviors that can transform a player's gaming experience. However, deploying RL agents in production environments presents two key challenges: (1) designing an effective reward function typically requires an RL expert, and (2) when a game's content or mechanics are modified, previously tuned reward weights may no longer be optimal. Towards the latter challenge, we prop…

Brain Network Analysis Based on Fine-tuned Self-supervised Model for Brain Disease Diagnosis
Yifei Tang, Hongjie Jiang, Changhong Jing, Hieu Pham, Shuqiang Wang
https://arxiv.org/abs/2506.11671

Brain Network Analysis Based on Fine-tuned Self-supervised Model for Brain Disease Diagnosis
Functional brain network analysis has become an indispensable tool for brain disease analysis. It is profoundly impacted by deep learning methods, which can characterize complex connections between ROIs. However, the research on foundation models of brain network is limited and constrained to a single dimension, which restricts their extensive application in neuroscience. In this study, we propose a fine-tuned brain network model for brain disease diagnosis. It expands brain region representati…

Mmm. My TV (tuned to algorithmic YouTube) is showing me the #BeeGees performing a lengthy medley of #Beatles tracks. It's nice. Surprised I haven't seen it before.

Emergent Quasiparticles \& Field-Tuned RIXS Spectra in a Trimerized Spin-1/2 Chain
Subhajyoti Pal, Pradeep Thakur, Ashis Kumar Nandy, Anamitra Mukherjee
https://arxiv.org/abs/2505.23208

Emergent Quasiparticles \& Field-Tuned RIXS Spectra in a Trimerized Spin-1/2 Chain
We investigate spin-flip excitations in the spin-1/2 trimer chain $\rm{Cu_3(P_2O_6OH)_2}$, featuring an antiferromagnetic exchange motif $J_1$-$J_1$-$J_2$ with $J_1 < J_2$. Using density matrix renormalization group (DMRG) simulations, we demonstrate that single-spin-flip processes induced by resonant inelastic X-ray scattering (RIXS) generate emergent gapless modes governed by the underlying trimer periodicity alongside distinct high-energy excitations. By combining exact diagonalization and r…

This https://arxiv.org/abs/2207.01666 has been replaced.
link: https://scholar.google.com/scholar?q=a

Cutoff stability of multivariate geometric Brownian motion
This article establishes cutoff convergence or abrupt convergence of three statistical quantities for multivariate (Hurwitz) stable geometric Brownian motion: the autocorrelation function, the Wasserstein distance between the current state and its degenerate limiting measure, and, finally, anti-concentration probabilities, which yield a fine-tuned trade-off between almost sure rates and the respective integrability of the random modulus of convergence using a quantitative Borel--Cantelli Lemma.…

This https://arxiv.org/abs/2505.20720 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csNE_…

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Floating hybrid wind-wave systems combine offshore wind platforms with wave energy converters (WECs) to create cost-effective and reliable energy solutions. Adequately designed and tuned WECs are essential to avoid unwanted loads disrupting turbine motion while efficiently harvesting wave energy. These systems diversify energy sources, enhancing energy security and reducing supply risks while providing a more consistent power output by smoothing energy production variability. However, optimisin…

Waiting for my new #patreon exclusive video to render! I'm very excited to post it.
Stay Tuned at https://www.patreon.com/raysofred

Can boundary configuration be tuned to optimize directional quantum steering harvesting?
Xiao-Li Huang, Xiao-Ying Jiang, Yu-Xuan Wang, Si-Yu Liu, Zejun Wang, Shu-Min Wu
https://arxiv.org/abs/2506.18734

Can boundary configuration be tuned to optimize directional quantum steering harvesting?
We investigate the harvesting of quantum steering and its asymmetry between two static detectors locally interacting with a vacuum massless scalar field near an infinite, perfectly reflecting boundary. The detectors are arranged either parallel or orthogonal to the boundary, with detector $B$ assumed to have an energy gap greater than or equal to that of detector $A$. It is interesting to observe that, with increasing distance between the detectors and the boundary, the boundary tends to suppre…

Ticket sales for #NixCon2025 are open! 🎉 🎆
https://discourse.nixos.org/t/ticket-sales-started-for-nixcon-2025/64756

SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition
Muhammad Umar Farooq, Oscar Saz
https://arxiv.org/abs/2506.22143

SAGE: Spliced-Audio Generated Data for Enhancing Foundational Models in Low-Resource Arabic-English Code-Switched Speech Recognition
This paper investigates the performance of various speech SSL models on dialectal Arabic (DA) and Arabic-English code-switched (CS) speech. To address data scarcity, a modified audio-splicing approach is introduced to generate artificial CS speech data. Fine-tuning an already fine-tuned SSL model with the proposed Spliced-Audio Generated (SAGE) data results in an absolute improvement on Word Error Rate (WER) of 7.8% on Arabic and English CS benchmarks. Additionally, an Experience Replay (ER) in…

Mixture of Encoders is a vector-native alternative that models both structured and unstructured data in a unified embedding space. Join Filip Makraduli as he introduces the method, demonstrates how it powers natural language search and real-time recommendations, and shares open-source tools and benchmarks for replacing complex hybrid stacks.
Learn more:

Crystal field tuned spin-flip luminescence in NiPS3
L\'eonard Schue, Nashra Pistawala, Hebatalla Elnaggar, Yannick Klein, Christophe Bellin, Johan Biscaras, Fausto Sirotti, Yves Lassailly, Fabian Cadiz, Luminita Harnagea, Abhay Shukla
https://arxiv.org/abs/2506.13381

Crystal field tuned spin-flip luminescence in NiPS3
Layered magnetic materials potentially hold the key to future applications based on optical control and manipulation of magnetism. NiPS3, a prototype member of this family, is antiferromagnetic below 155 K and exhibits sharp photoluminescence associated to a transition between a triplet ground state and a singlet excited state. The nature of the luminescent transition is a matter of current debate and so is an eventual fundamental link of this excitation to magnetism. Here we provide answers th…

Instruction-Tuned Video-Audio Models Elucidate Functional Specialization in the Brain
Subba Reddy Oota, Khushbu Pahwa, Prachi Jindal, Satya Sai Srinath Namburi, Maneesh Singh, Tanmoy Chakraborty, Bapi S. Raju, Manish Gupta
https://arxiv.org/abs/2506.08277

Instruction-Tuned Video-Audio Models Elucidate Functional Specialization in the Brain
Recent voxel-wise multimodal brain encoding studies have shown that multimodal large language models (MLLMs) exhibit a higher degree of brain alignment compared to unimodal models in both unimodal and multimodal stimulus settings. More recently, instruction-tuned multimodal models have shown to generate task-specific representations that align strongly with brain activity. However, prior work evaluating the brain alignment of MLLMs has primarily focused on unimodal settings or relied on non-ins…

Tunable bound states in the continuum through hybridization of 1D and 2D metasurfaces
Fedor Kovalev, Mariusz Martyniuk, Andrey Miroshnichenko, Ilya Shadrivov
https://arxiv.org/abs/2506.21147

Tunable bound states in the continuum through hybridization of 1D and 2D metasurfaces
This work presents a novel approach to create and dynamically control quasi-Bound States in the Continuum (BIC) resonances through the hybridization of 1D and 2D metasurfaces using micro-electromechanical systems (MEMS). By introducing out-of-plane symmetry breaking through a silicon MEMS membrane positioned above a 1D silicon metasurface, the quasi-BIC resonance's central wavelength and quality factor are precisely tuned. The proposed design achieves ultranarrow resonance linewidths with the s…

One place I don't mind algorithms is in music apps. I just got Weird Al, Korn, Tiesto, Lynyrd Skynyrd, and Dr. Dre. That's one well tuned algorithm.

HLTCOE at LiveRAG: GPT-Researcher using ColBERT retrieval
Kevin Duh, Eugene Yang, Orion Weller, Andrew Yates, Dawn Lawrie
https://arxiv.org/abs/2506.22356 …

HLTCOE at LiveRAG: GPT-Researcher using ColBERT retrieval
The HLTCOE LiveRAG submission utilized the GPT-researcher framework for researching the context of the question, filtering the returned results, and generating the final answer. The retrieval system was a ColBERT bi-encoder architecture, which represents a passage with many dense tokens. Retrieval used a local, compressed index of the FineWeb10-BT collection created with PLAID-X, using a model fine-tuned for multilingual retrieval. Query generation from context was done with Qwen2.5-7B-Instruct…

Resonances of recurrence time of monitored quantum walks
Ruoyu Yin, Qingyuan Wang, Sabine Tornow, Eli Barkai
https://arxiv.org/abs/2506.19832 https://

Resonances of recurrence time of monitored quantum walks
The recurrence time is the time a process first returns to its initial state. Using quantum walks on a graph, the recurrence time is defined through stroboscopic monitoring of the arrival of the particle to a node of the system. When the time interval between repeated measurements is tuned in such a way that eigenvalues of the unitary become degenerate, the mean recurrence time exhibits resonances. These resonances imply faster mean recurrence times, which were recorded on quantum computers. Th…

This https://arxiv.org/abs/2502.09930 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_qu…

Long-Lived Photon Blockade with Weak Optical Nonlinearity
In conventional photon blockade, the occupation of a cavity mode by more than one photon is suppressed via strong optical nonlinearity. An alternative, called unconventional photon blockade, can occur under weak nonlinearity by relying on quantum interference between fine-tuned cavities. A serious limitation is the very short antibunching time window, orders of magnitude less than the cavity lifetime. We present a method to achieve photon blockade over a large time window of several cavity life…

This https://arxiv.org/abs/2505.20720 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csNE_…

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Floating hybrid wind-wave systems combine offshore wind platforms with wave energy converters (WECs) to create cost-effective and reliable energy solutions. Adequately designed and tuned WECs are essential to avoid unwanted loads disrupting turbine motion while efficiently harvesting wave energy. These systems diversify energy sources, enhancing energy security and reducing supply risks while providing a more consistent power output by smoothing energy production variability. However, optimisin…

Comment on "Long-range crossed Andreev reflection in a topological insulator nanowire proximitized by a superconductor" by Junya Feng et al
E. S. Tikhonov, V. S. Khrapai
https://arxiv.org/abs/2505.23490

Comment on "Long-range crossed Andreev reflection in a topological insulator nanowire proximitized by a superconductor" by Junya Feng et al
We argue that the interpretation of the experiment [Nature Physics 21, 708-715 (2025)] is misleading in two respects. First, the bias voltages impact the non-local differential conductance randomly, rather than systematically, and the bias symmetry of the non-local conductance in Fig. 3 can be explained by a fine tuned self-gating effect. Second, the full knowledge of the conductance matrix is insufficient to conclude on the relative values of the crossed-Andreev and elastic cotunneling probabi…

Day 13 (oh, really? ;))
Registration Form Implementation
I've just finished implementing a registration form with validation and language switching using Next.js and React Hook Form. Now users can register with dynamic language support (English/Polish) and data validation (email, password, phone).
Unfortunately, my account on Write.as has been temporarily blocked, so details about the implementation will be available once the account is unlocked. Stay tuned! 😊

PhantomHunter: Detecting Unseen Privately-Tuned LLM-Generated Text via Family-Aware Learning
Yuhui Shi, Yehan Yang, Qiang Sheng, Hao Mi, Beizhe Hu, Chaoxi Xu, Juan Cao
https://arxiv.org/abs/2506.15683

PhantomHunter: Detecting Unseen Privately-Tuned LLM-Generated Text via Family-Aware Learning
With the popularity of large language models (LLMs), undesirable societal problems like misinformation production and academic misconduct have been more severe, making LLM-generated text detection now of unprecedented importance. Although existing methods have made remarkable progress, a new challenge posed by text from privately tuned LLMs remains underexplored. Users could easily possess private LLMs by fine-tuning an open-source one with private corpora, resulting in a significant performanc…

Effect of Vaccine Dose Intervals: Considering Immunity Levels, Vaccine Efficacy, and Strain Variants for Disease Control Strategy
Samiran Ghosh, Malay Banerjee, Amit K Chattopadhyay
https://arxiv.org/abs/2505.21132

Effect of Vaccine Dose Intervals: Considering Immunity Levels, Vaccine Efficacy, and Strain Variants for Disease Control Strategy
In this study, we present an immuno-epidemic model to understand mitigation options during an epidemic break. The model incorporates comorbidity and multiple-vaccine doses through a system of coupled integro-differential equations to analyze the epidemic rate and intensity from a knowledge of the basic reproduction number and time-distributed rate functions. Our modeling results show that the interval between vaccine doses is a key control parameter that can be tuned to significantly influence …

Foundation Artificial Intelligence Models for Health Recognition Using Face Photographs (FAHR-Face)
Fridolin Haugg, Grace Lee, John He, Leonard N\"urnberg, Dennis Bontempi, Danielle S. Bitterman, Paul Catalano, Vasco Prudente, Dmitrii Glubokov, Andrew Warrington, Suraj Pai, Dirk De Ruysscher, Christian Guthier, Benjamin H. Kann, Vadim N. Gladyshev, Hugo JWL Aerts, Raymond H. Mak

Foundation Artificial Intelligence Models for Health Recognition Using Face Photographs (FAHR-Face)
Background: Facial appearance offers a noninvasive window into health. We built FAHR-Face, a foundation model trained on >40 million facial images and fine-tuned it for two distinct tasks: biological age estimation (FAHR-FaceAge) and survival risk prediction (FAHR-FaceSurvival). Methods: FAHR-FaceAge underwent a two-stage, age-balanced fine-tuning on 749,935 public images; FAHR-FaceSurvival was fine-tuned on 34,389 photos of cancer patients. Model robustness (cosmetic surgery, makeup, pose, l…

I was coming back home in the bus and every house that had a TV near their windows were tuned to the FP2 😅 #CanadianGP

This https://arxiv.org/abs/2505.18700 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCV_…

GRE Suite: Geo-localization Inference via Fine-Tuned Vision-Language Models and Enhanced Reasoning Chains
Recent advances in Visual Language Models (VLMs) have demonstrated exceptional performance in visual reasoning tasks. However, geo-localization presents unique challenges, requiring the extraction of multigranular visual cues from images and their integration with external world knowledge for systematic reasoning. Current approaches to geo-localization tasks often lack robust reasoning mechanisms and explainability, limiting their effectiveness. To address these limitations, we propose the Geo …

Modeling phase transformations in Mn-rich disordered rocksalt cathodes with charge-informed machine-learning interatomic potentials
Peichen Zhong, Bowen Deng, Shashwat Anand, Tara Mishra, Gerbrand Ceder
https://arxiv.org/abs/2506.20605

Modeling phase transformations in Mn-rich disordered rocksalt cathodes with charge-informed machine-learning interatomic potentials
Mn-rich disordered rocksalt (DRX) cathode materials exhibit a phase transformation from a disordered to a partially disordered spinel-like structure ($δ$-phase) during electrochemical cycling. In this computational study, we used charge-informed molecular dynamics with a fine-tuned CHGNet foundation potential to investigate the phase transformation in Li$_{x}$Mn$_{0.8}$Ti$_{0.1}$O$_{1.9}$F$_{0.1}$. Our results indicate that transition metal migration occurs and reorders to form the spinel-like…

Correlating instruction-tuning (in multimodal models) with vision-language processing (in the brain)
Subba Reddy Oota, Akshett Jindal, Ishani Mondal, Khushbu Pahwa, Satya Sai Srinath Namburi, Manish Shrivastava, Maneesh Singh, Bapi S. Raju, Manish Gupta
https://arxiv.org/abs/2505.20029

Correlating instruction-tuning (in multimodal models) with vision-language processing (in the brain)
Transformer-based language models, though not explicitly trained to mimic brain recordings, have demonstrated surprising alignment with brain activity. Progress in these models-through increased size, instruction-tuning, and multimodality-has led to better representational alignment with neural data. Recently, a new class of instruction-tuned multimodal LLMs (MLLMs) have emerged, showing remarkable zero-shot capabilities in open-ended multimodal vision tasks. However, it is unknown whether MLLM…

On the Practicability of Ceramic-Tiled Walls for Sound Absorption by Tuning Cavities
Ozgur T. Tugut, Brahim Lemkalli, Qingxiang Ji, Mahmoud Addouche, Benjamin Vial, S\'ebastien Guenneau, Richard Craster, Claudio Bizzaglia, Bogdan Ungureanu, Muamer Kadic
https://arxiv.org/abs/2506.17699…

On the Practicability of Ceramic-Tiled Walls for Sound Absorption by Tuning Cavities
We present the practicality of structuring ceramic tiles for enhancing sound absorption on rigid walls. The cornerstone of our methodology is to structure walls with cavities so that walls effectively behave as heterogeneous absorbing surfaces over a large frequency bandwidth. Using this approach, ceramic tiled walls are developed by integrating tuned cavity structures based on Helmholtz resonators. Such a design leverages the empty joints between tiles to form resonator necks, while the space …

FedShield-LLM: A Secure and Scalable Federated Fine-Tuned Large Language Model
Md Jueal Mia, M. Hadi Amini
https://arxiv.org/abs/2506.05640 https://…

FedShield-LLM: A Secure and Scalable Federated Fine-Tuned Large Language Model
Federated Learning (FL) offers a decentralized framework for training and fine-tuning Large Language Models (LLMs) by leveraging computational resources across organizations while keeping sensitive data on local devices. It addresses privacy and security concerns while navigating challenges associated with the substantial computational demands of LLMs, which can be prohibitive for small and medium-sized organizations. FL supports the development of task-specific LLMs for cross-silo applications…

DreamAnywhere: Object-Centric Panoramic 3D Scene Generation
Edoardo Alberto Dominici, Jozef Hladky, Floor Verhoeven, Lukas Radl, Thomas Deixelberger, Stefan Ainetter, Philipp Drescher, Stefan Hauswiesner, Arno Coomans, Giacomo Nazzaro, Konstantinos Vardis, Markus Steinberger
https://arxiv.org/abs/2506.20367

DreamAnywhere: Object-Centric Panoramic 3D Scene Generation
Recent advances in text-to-3D scene generation have demonstrated significant potential to transform content creation across multiple industries. Although the research community has made impressive progress in addressing the challenges of this complex task, existing methods often generate environments that are only front-facing, lack visual fidelity, exhibit limited scene understanding, and are typically fine-tuned for either indoor or outdoor settings. In this work, we address these issues and …

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Mehdi Neshat, Nataliia Y. Sergiienko, Leandro S. P. da Silva, Seyedali Mirjalili, Amir H. Gandomi, Ossama Abdelkhalik, John Boland
https://arxiv.org/abs/2505.20720

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Floating hybrid wind-wave systems combine offshore wind platforms with wave energy converters (WECs) to create cost-effective and reliable energy solutions. Adequately designed and tuned WECs are essential to avoid unwanted loads disrupting turbine motion while efficiently harvesting wave energy. These systems diversify energy sources, enhancing energy security and reducing supply risks while providing a more consistent power output by smoothing energy production variability. However, optimisin…

Evolution of charge correlations in the hole-doped kagome superconductor CsV$_{3-x}$Ti$_x$Sb$_5$
Ganesh Pokharel, Canxun Zhang, Evgeny Redekop, Brenden R. Ortiz, Andrea N. Capa Salinas, Sarah Schwarz, Steven J. Gomez Alvarado, Suchismita Sarker, Andrea F. Young, Stephen D. Wilson
https://arxiv.org/abs/2506.13941

Evolution of charge correlations in the hole-doped kagome superconductor CsV$_{3-x}$Ti$_x$Sb$_5$
The interplay between superconductivity and charge correlations in the kagome metal CsV$_3$Sb$_5$ can be tuned by external perturbations such as doping or pressure. Here we present a study of charge correlations and superconductivity upon hole doping via Ti substitution on the V kagome sites in CsV$_{3-x}$Ti$_x$Sb$_5$ via synchrotron x-ray diffraction and scanning SQUID measurements. While the superconducting phase, as viewed via the vortex structure, remains conventional and unchanged across t…

Cyclicity of interaction frame transformations
Michael C D Tayler, Mohamed Sabba
https://arxiv.org/abs/2506.20594 https://arxiv.org/p…

Cyclicity of interaction frame transformations
We identify a cyclic property of rotation sequences involving piecewise displacements $β$ about arbitrary axes in three dimensions. Specifically, when transformation to the toggling frame is applied successively $m$ times, for $β=2π/m$ the original sequence returns. This main result unites several families of rotation sequences used for error-tuned control across quantum technologies, from NMR and MRI to atomic clocks and atom-scale computing. For the widest class of cycle, $m=2$, we highlig…

SE-Merging: A Self-Enhanced Approach for Dynamic Model Merging
Zijun Chen, Zhanpeng Zhou, Bo Zhang, Weinan Zhang, Xi Sun, Junchi Yan
https://arxiv.org/abs/2506.18135

SE-Merging: A Self-Enhanced Approach for Dynamic Model Merging
Model merging has gained increasing attention due to its intriguing property: interpolating the parameters of different task-specific fine-tuned models leads to multi-task abilities. However, despite its empirical success, the underlying mechanisms of model merging remain poorly understood. In this work, we delve into the mechanism behind model merging from a representation perspective. Our analysis reveals that model merging achieves multi-task abilities through two key capabilities: i) distin…

ScaleRTL: Scaling LLMs with Reasoning Data and Test-Time Compute for Accurate RTL Code Generation
Chenhui Deng, Yun-Da Tsai, Guan-Ting Liu, Zhongzhi Yu, Haoxing Ren
https://arxiv.org/abs/2506.05566

ScaleRTL: Scaling LLMs with Reasoning Data and Test-Time Compute for Accurate RTL Code Generation
Recent advances in large language models (LLMs) have enabled near-human performance on software coding benchmarks, but their effectiveness in RTL code generation remains limited due to the scarcity of high-quality training data. While prior efforts have fine-tuned LLMs for RTL tasks, they do not fundamentally overcome the data bottleneck and lack support for test-time scaling due to their non-reasoning nature. In this work, we introduce ScaleRTL, the first reasoning LLM for RTL coding that scal…

Can Common VLMs Rival Medical VLMs? Evaluation and Strategic Insights
Yuan Zhong, Ruinan Jin, Xiaoxiao Li, Qi Dou
https://arxiv.org/abs/2506.17337 https://…

Can Common VLMs Rival Medical VLMs? Evaluation and Strategic Insights
Medical vision-language models (VLMs) leverage large-scale pretraining for diverse imaging tasks but require substantial computational and data resources. Meanwhile, common or general-purpose VLMs (e.g., CLIP, LLaVA), though not trained for medical use, show promise with fine-tuning. This raises a key question: Can efficient fine-tuned common VLMs rival generalist medical VLMs for solving specific medical imaging tasks? This study systematically evaluates common and medical VLMs across disease …

This https://arxiv.org/abs/2505.06369 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_hept…

Yang-Lee Quantum Criticality in Various Dimensions
The Yang-Lee universality class arises when imaginary magnetic field is tuned to its critical value in the paramagnetic phase of the $d<6$ Ising model. In $d=2$, this non-unitary Conformal Field Theory (CFT) is exactly solvable via the $M(2,5)$ minimal model. As found long ago by von Gehlen using Exact Diagonalization, the corresponding real-time, quantum critical behavior arises in the periodic Ising spin chain when the imaginary longitudinal magnetic field is tuned to its critical value from …

Join Kevin Liang at this year's Berlin Buzzwords, where he will discuss how Apache Solr/Lucene builds dense vector indexes and talk about how he and his team optimised their dense vector setup, sharing the challenges they faced and the best practices they learned along the way.
Learn more: https…

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Mehdi Neshat, Nataliia Y. Sergiienko, Leandro S. P. da Silva, Seyedali Mirjalili, Amir H. Gandomi, Ossama Abdelkhalik, John Boland
https://arxiv.org/abs/2505.20720

Hybrid Wave-wind System Power Optimisation Using Effective Ensemble Covariance Matrix Adaptation Evolutionary Algorithm
Floating hybrid wind-wave systems combine offshore wind platforms with wave energy converters (WECs) to create cost-effective and reliable energy solutions. Adequately designed and tuned WECs are essential to avoid unwanted loads disrupting turbine motion while efficiently harvesting wave energy. These systems diversify energy sources, enhancing energy security and reducing supply risks while providing a more consistent power output by smoothing energy production variability. However, optimisin…

Slimmer Geminals For Accurate F12 Electronic Structure Models
Samuel R. Powell, Edward F. Valeev
https://arxiv.org/abs/2506.10780 https://

Slimmer Geminals For Accurate F12 Electronic Structure Models
The Slater-type F12 geminal lengthscales originally tuned for the second-order Møller-Plesset F12 method are too large for higher-order F12 methods formulated using the SP (diagonal fixed-coefficient spin-adapted) F12 ansatz. The new geminal parameters reported herein reduce the basis set incompleteness errors (BSIE) of absolute coupled-cluster singles and doubles F12 correlation energies by a significant - and increasing with the cardinal number of the basis - margin. The effect of geminal re…

Plasmonic properties of individual bismuth nanoparticles
Michael Folt\'yn, Michal Kvapil, Tom\'a\v{s} \v{S}ikola, Michal Hor\'ak
https://arxiv.org/abs/2506.16248

Plasmonic properties of individual bismuth nanoparticles
Bismuth nanoparticles are being investigated due to their reported photothermal and photocatalytic properties. In this study, we synthesised spherical bismuth nanoparticles (50-600 nm) and investigated their structural and optical properties at the single particle level using analytical transmission electron microscopy. Our experimental results, supported by numerical simulations, demonstrate that bismuth nanoparticles support localised surface plasmon resonances, which can be tuned from the ne…

When Detection Fails: The Power of Fine-Tuned Models to Generate Human-Like Social Media Text
Hillary Dawkins, Kathleen C. Fraser, Svetlana Kiritchenko
https://arxiv.org/abs/2506.09975

When Detection Fails: The Power of Fine-Tuned Models to Generate Human-Like Social Media Text
Detecting AI-generated text is a difficult problem to begin with; detecting AI-generated text on social media is made even more difficult due to the short text length and informal, idiosyncratic language of the internet. It is nonetheless important to tackle this problem, as social media represents a significant attack vector in online influence campaigns, which may be bolstered through the use of mass-produced AI-generated posts supporting (or opposing) particular policies, decisions, or event…

Private Continual Counting of Unbounded Streams
Ben Jacobsen, Kassem Fawaz
https://arxiv.org/abs/2506.15018 https://arxiv.org/pdf/250…

Private Continual Counting of Unbounded Streams
We study the problem of differentially private continual counting in the unbounded setting where the input size $n$ is not known in advance. Current state-of-the-art algorithms based on optimal instantiations of the matrix mechanism cannot be directly applied here because their privacy guarantees only hold when key parameters are tuned to $n$. Using the common `doubling trick' avoids knowledge of $n$ but leads to suboptimal and non-smooth error. We solve this problem by introducing novel matrix…

Data Driven Approach to Input Shaping for Vibration Suppression in a Flexible Robot Arm
Jarkko Kotaniemi, Janne Saukkoriipi, Shuai Li, Markku Suomalainen
https://arxiv.org/abs/2506.14405

Data Driven Approach to Input Shaping for Vibration Suppression in a Flexible Robot Arm
This paper presents a simple and effective method for setting parameters for an input shaper to suppress the residual vibrations in flexible robot arms using a data-driven approach. The parameters are adaptively tuned in the workspace of the robot by interpolating previously measured data of the robot's residual vibrations. Input shaping is a simple and robust technique to generate vibration-reduced shaped commands by a convolution of an impulse sequence with the desired input command. The gene…

A Study on the Fine-Tuning Performance of Universal Machine-Learned Interatomic Potentials (U-MLIPs)
Xiaoqing Liu, Kehan Zeng, Yangshuai Wang, Teng Zhao
https://arxiv.org/abs/2506.07401

A Study on the Fine-Tuning Performance of Universal Machine-Learned Interatomic Potentials (U-MLIPs)
Universal machine-learned interatomic potentials (U-MLIPs) have demonstrated effectiveness across diverse atomistic systems but often require fine-tuning for task-specific accuracy. We investigate the fine-tuning of two MACE-based foundation models, MACE-MP-0 and its variant MACE-MP-0b, and identify key insights. Fine-tuning on task-specific datasets enhances accuracy and, in some cases, outperforms models trained from scratch. Additionally, fine-tuned models benefit from faster convergence due…

Evaluating the Performance and Efficiency of Sentence-BERT for Code Comment Classification
Fabian C. Pe\~na, Steffen Herbold
https://arxiv.org/abs/2506.08581

Evaluating the Performance and Efficiency of Sentence-BERT for Code Comment Classification
This work evaluates Sentence-BERT for a multi-label code comment classification task seeking to maximize the classification performance while controlling efficiency constraints during inference. Using a dataset of 13,216 labeled comment sentences, Sentence-BERT models are fine-tuned and combined with different classification heads to recognize comment types. While larger models outperform smaller ones in terms of F1, the latter offer outstanding efficiency, both in runtime and GFLOPS. As result…

State-Space Models in Efficient Whispered and Multi-dialect Speech Recognition
Aref Farhadipour, Homayoon Beigi, Volker Dellwo, Hadi Veisi
https://arxiv.org/abs/2506.16969

State-Space Models in Efficient Whispered and Multi-dialect Speech Recognition
Whispered speech recognition presents significant challenges for conventional automatic speech recognition systems, particularly when combined with dialect variation. However, utilizing an efficient method to solve this problem using a low-range dataset and processing load is beneficial. This paper proposes a solution using a Mamba-based state-space model and four fine-tuned self-supervised models consisting of Wav2Vec2, WavLM, HuBERT, and Whisper to address the dual challenges of whispered spe…

I really appreciate in how bad Amazon is at recommendations in 2025, when everything else is headed for some kind of AI fine-tuned personalization. Meanwhile Amazon is like "I see you bought an allen wrench last month, I bet you want even more allen wrenches"

X-Shooting ULLYSES: Massive stars at low metallicity. XIII. Putting the bi-stability jump to the test in the LMC
T. Alkousa (Astrophysics Research Cluster, School of Mathematical and Physical Sciences, University of Sheffield), P. A. Crowther (Astrophysics Research Cluster, School of Mathematical and Physical Sciences, University of Sheffield), J. M. Bestenlehner (Astrophysics Research Cluster, School of Mathematical and Physical Sciences, University of Sheffield, School of Chemical, M…

X-Shooting ULLYSES: Massive stars at low metallicity. XIII. Putting the bi-stability jump to the test in the LMC
We aim to investigate the theoretical bi-stability jump, which predicts an increase in mass-loss rates below 21 kK. We further aim to constrain the photospheric and wind parameters of a sample of 16 LMC late-O and B supergiants. We utilise the 1D, non-LTE radiative transfer model CMFGEN in a grid-based approach and subsequent fine-tuned spectroscopic fitting procedure to determine the stellar and wind parameters of each star. We apply this method to ultra-violet data from the ULLYSES programme …

Screen Hijack: Visual Poisoning of VLM Agents in Mobile Environments
Xuan Wang, Siyuan Liang, Zhe Liu, Yi Yu, Yuliang Lu, Xiaochun Cao, Ee-Chien Chang
https://arxiv.org/abs/2506.13205

Screen Hijack: Visual Poisoning of VLM Agents in Mobile Environments
With the growing integration of vision-language models (VLMs), mobile agents are now widely used for tasks like UI automation and camera-based user assistance. These agents are often fine-tuned on limited user-generated datasets, leaving them vulnerable to covert threats during the training process. In this work we present GHOST, the first clean-label backdoor attack specifically designed for mobile agents built upon VLMs. Our method manipulates only the visual inputs of a portion of the traini…

Anomaly, Class Division, and Decoupling in Wealth Dynamics
Jaeseok Hur, Meesoon Ha, Hawoong Jeong
https://arxiv.org/abs/2506.08175 https://

Anomaly, Class Division, and Decoupling in Wealth Dynamics
Economic inequality is shaped by the agent-network structure, the interaction between agents, and the individual agent's ability. We provide a comprehensive picture of anomalous diffusion, economic class division, and bimodal wealth distribution in an agent-based model, where the allocation of heterogeneous agent abilities/growth rates is tuned in sparse regular networks.In particular, we focus on the statistical characteristics of logarithmic scaled normalized wealth distributions with two abi…

Screen Hijack: Visual Poisoning of VLM Agents in Mobile Environments
Xuan Wang, Siyuan Liang, Zhe Liu, Yi Yu, Yuliang Lu, Xiaochun Cao, Ee-Chien Chang
https://arxiv.org/abs/2506.13205

Screen Hijack: Visual Poisoning of VLM Agents in Mobile Environments
With the growing integration of vision-language models (VLMs), mobile agents are now widely used for tasks like UI automation and camera-based user assistance. These agents are often fine-tuned on limited user-generated datasets, leaving them vulnerable to covert threats during the training process. In this work we present GHOST, the first clean-label backdoor attack specifically designed for mobile agents built upon VLMs. Our method manipulates only the visual inputs of a portion of the traini…

Passing the Turing Test in Political Discourse: Fine-Tuning LLMs to Mimic Polarized Social Media Comments
. Pazzaglia, V. Vendetti, L. D. Comencini, F. Deriu, V. Modugno
https://arxiv.org/abs/2506.14645

Passing the Turing Test in Political Discourse: Fine-Tuning LLMs to Mimic Polarized Social Media Comments
The increasing sophistication of large language models (LLMs) has sparked growing concerns regarding their potential role in exacerbating ideological polarization through the automated generation of persuasive and biased content. This study explores the extent to which fine-tuned LLMs can replicate and amplify polarizing discourse within online environments. Using a curated dataset of politically charged discussions extracted from Reddit, we fine-tune an open-source LLM to produce context-aware…

Tunable corner states in topological insulators with long-range hoppings and diverse shapes
Fang Qin, Rui Chen
https://arxiv.org/abs/2506.12467 https://

Tunable corner states in topological insulators with long-range hoppings and diverse shapes
In this work, we develop a theoretical framework for the control of corner modes in higher-order topological insulators (HOTIs) featuring long-range hoppings and diverse geometries, enabling precise tunability of their spatial positions. First, we demonstrate that the locations of corner states can be finely tuned by varying long-range hoppings in a circular HOTI, as revealed by a detailed edge theory analysis and the condition of vanishing Dirac mass. Moreover, we show that long-range hoppings…

This https://arxiv.org/abs/2311.16593 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_ees…

Empowering COVID-19 Detection: Optimizing Performance Through Fine-Tuned EfficientNet Deep Learning Architecture
The worldwide COVID-19 pandemic has profoundly influenced the health and everyday experiences of individuals across the planet. It is a highly contagious respiratory disease requiring early and accurate detection to curb its rapid transmission. Initial testing methods primarily revolved around identifying the genetic composition of the coronavirus, exhibiting a relatively low detection rate and requiring a time-intensive procedure. To address this challenge, experts have suggested using radiolo…

Evaluating Prompt-Driven Chinese Large Language Models: The Influence of Persona Assignment on Stereotypes and Safeguards
Geng Liu, Li Feng, Carlo Alberto Bono, Songbo Yang, Mengxiao Zhu, Francesco Pierri
https://arxiv.org/abs/2506.04975

Evaluating Prompt-Driven Chinese Large Language Models: The Influence of Persona Assignment on Stereotypes and Safeguards
Recent research has highlighted that assigning specific personas to large language models (LLMs) can significantly increase harmful content generation. Yet, limited attention has been given to persona-driven toxicity in non-Western contexts, particularly in Chinese-based LLMs. In this paper, we perform a large-scale, systematic analysis of how persona assignment influences refusal behavior and response toxicity in Qwen, a widely-used Chinese language model. Utilizing fine-tuned BERT classifiers…

Training-free LLM Merging for Multi-task Learning
Zichuan Fu, Xian Wu, Yejing Wang, Wanyu Wang, Shanshan Ye, Hongzhi Yin, Yi Chang, Yefeng Zheng, Xiangyu Zhao
https://arxiv.org/abs/2506.12379

Training-free LLM Merging for Multi-task Learning
Large Language Models (LLMs) have demonstrated exceptional capabilities across diverse natural language processing (NLP) tasks. The release of open-source LLMs like LLaMA and Qwen has triggered the development of numerous fine-tuned models tailored for various tasks and languages. In this paper, we explore an important question: is it possible to combine these specialized models to create a unified model with multi-task capabilities. We introduces Hierarchical Iterative Merging (Hi-Merging), a …

Tunability of Robust Exciton-Trion Polaritons in Atomically Thin WS2 Monolayers
Xuguang Cao, Debao Zhang, Ji Zhou, Wanggui Ye, Changcheng Zheng, Kenji Watanabe, Takashi Taniguchi, Jiqiang Ning, Shijie Xu
https://arxiv.org/abs/2506.07030

Tunability of Robust Exciton-Trion Polaritons in Atomically Thin WS2 Monolayers
Herein, we present an experimental demonstration of the robust exciton-trion polaritons (ETPs) by measuring and simulating the resonance reflectance spectra of various configurational WS2 monolayers with different dielectric screenings. Moreover, the oscillator strength and decoherent behavior of such hybrid ETPs can be tuned via utilizing dielectric screening effect. The effect is attributed to the regulation of the Coulomb coupling between excitons and trions by changing the surrounding diele…

Serendipitous Recommendation with Multimodal LLM
Haoting Wang, Jianling Wang, Hao Li, Fangjun Yi, Mengyu Fu, Youwei Zhang, Yifan Liu, Liang Liu, Minmin Chen, Ed H. Chi, Lichan Hong, Haokai Lu
https://arxiv.org/abs/2506.08283

Serendipitous Recommendation with Multimodal LLM
Conventional recommendation systems succeed in identifying relevant content but often fail to provide users with surprising or novel items. Multimodal Large Language Models (MLLMs) possess the world knowledge and multimodal understanding needed for serendipity, but their integration into billion-item-scale platforms presents significant challenges. In this paper, we propose a novel hierarchical framework where fine-tuned MLLMs provide high-level guidance to conventional recommendation models, s…

Analog dual classifier via a time-modulated neuromorphic metasurface
M. Mousa, M. Moghaddaszadeh, M. Nouh
https://arxiv.org/abs/2506.04629 https://

Analog dual classifier via a time-modulated neuromorphic metasurface
A neuromorphic metasurface embodies mechanical intelligence by realizing physical neural architectures. It exploits guided wave scattering to conduct computations in an analog manner. Through multiple tuned waveguides, the neuromorphic system recognizes the features of an input signal and self-identifies its classification label. The computational input is introduced to the system through mechanical excitations at one edge, generating elastic waves which traverse multiple layers of resonant met…

BUT System for the MLC-SLM Challenge
Alexander Polok, Jiangyu Han, Dominik Klement, Samuele Cornell, Jan \v{C}ernock\'y, Luk\'a\v{s} Burget
https://arxiv.org/abs/2506.13414

BUT System for the MLC-SLM Challenge
We present a two-speaker automatic speech recognition (ASR) system that combines DiCoW -- a diarization-conditioned variant of Whisper -- with DiariZen, a diarization pipeline built on top of Pyannote. We first evaluate both systems in out-of-domain (OOD) multilingual scenarios without any fine-tuning. In this scenario, DiariZen consistently outperforms the baseline Pyannote diarization model, demonstrating strong generalization. Despite being fine-tuned on English-only data for target-speaker …

In-situ tuning of cavity dissipation and a topological transition in an atom-nanotip-cavity system
Taegyu Ha, Kiyanoush Goudarzi, Dowon Lee, Donggeon Kim, Eunchul Jeong, Uijin Kim, Myunghun Kim, Jinuk Kim, Moonjoo Lee
https://arxiv.org/abs/2506.07800

In-situ tuning of cavity dissipation and a topological transition in an atom-nanotip-cavity system
We theoretically investigate a method for tuning a dissipation rate in an atom-cavity system. By inserting a nanotip into the cavity mode, we estimate that the cavity dissipation rate increases by a factor of approximately 20, due to the scattering loss at the tip. As applications of our in-situ technique, we demonstrate that an exceptional line can be obtained when a single atom or ion is coupled to the resonator. Moreover, the position of an exceptional point is tuned by adjusting the decay r…

Exploring Cultural Variations in Moral Judgments with Large Language Models
Hadi Mohammadi, Efthymia Papadopoulou, Yasmeen F. S. S. Meijer, Ayoub Bagheri
https://arxiv.org/abs/2506.12433

Exploring Cultural Variations in Moral Judgments with Large Language Models
Large Language Models (LLMs) have shown strong performance across many tasks, but their ability to capture culturally diverse moral values remains unclear. In this paper, we examine whether LLMs can mirror variations in moral attitudes reported by two major cross-cultural surveys: the World Values Survey and the PEW Research Center's Global Attitudes Survey. We compare smaller, monolingual, and multilingual models (GPT-2, OPT, BLOOMZ, and Qwen) with more recent instruction-tuned models (GPT-4o,…

Tunable Epsilon Near Zero Metamaterial with Rotating Obround-Shaped Meta-Atoms
Rustam Balafendiev, Gagandeep Kaur, Jim A. Enriquez, Gaganpreet Singh, Alexander J. Millar, Jon E. Gudmundsson, Pavel Belov
https://arxiv.org/abs/2506.04428

Tunable Epsilon Near Zero Metamaterial with Rotating Obround-Shaped Meta-Atoms
A new design of a microwave-range ENZ metamaterial consisting of rods with an obround cross-section is proposed. The plasma frequency of the metamaterial can be tuned by rotating the constituent meta-atoms. Tunability of the plasma frequency by 26% is demonstrated both experimentally and numerically. The observed tuning range is dramatically higher than in the one observed in natural materials at optical range.

SOFT: Selective Data Obfuscation for Protecting LLM Fine-tuning against Membership Inference Attacks
Kaiyuan Zhang, Siyuan Cheng, Hanxi Guo, Yuetian Chen, Zian Su, Shengwei An, Yuntao Du, Charles Fleming, Ashish Kundu, Xiangyu Zhang, Ninghui Li
https://arxiv.org/abs/2506.10424

SOFT: Selective Data Obfuscation for Protecting LLM Fine-tuning against Membership Inference Attacks
Large language models (LLMs) have achieved remarkable success and are widely adopted for diverse applications. However, fine-tuning these models often involves private or sensitive information, raising critical privacy concerns. In this work, we conduct the first comprehensive study evaluating the vulnerability of fine-tuned LLMs to membership inference attacks (MIAs). Our empirical analysis demonstrates that MIAs exploit the loss reduction during fine-tuning, making them highly effective in re…

Group then Scale: Dynamic Mixture-of-Experts Multilingual Language Model
Chong Li, Yingzhuo Deng, Jiajun Zhang, Chengqing Zong
https://arxiv.org/abs/2506.12388

Group then Scale: Dynamic Mixture-of-Experts Multilingual Language Model
The curse of multilinguality phenomenon is a fundamental problem of multilingual Large Language Models (LLMs), where the competition between massive languages results in inferior performance. It mainly comes from limited capacity and negative transfer between dissimilar languages. To address this issue, we propose a method to dynamically group and scale up the parameters of multilingual LLM while boosting positive transfer among similar languages. Specifically, the model is first tuned on monol…

FinBERT2: A Specialized Bidirectional Encoder for Bridging the Gap in Finance-Specific Deployment of Large Language Models
Xuan Xu, Fufang Wen, Beilin Chu, Zhibing Fu, Qinhong Lin, Jiaqi Liu, Binjie Fei, Zhongliang Yang, Linna Zhou, Yu Li
https://arxiv.org/abs/2506.06335

FinBERT2: A Specialized Bidirectional Encoder for Bridging the Gap in Finance-Specific Deployment of Large Language Models
In natural language processing (NLP), the focus has shifted from encoder-only tiny language models like BERT to decoder-only large language models(LLMs) such as GPT-3. However, LLMs' practical application in the financial sector has revealed three limitations: (1) LLMs often perform worse than fine-tuned BERT on discriminative tasks despite costing much higher computational resources, such as market sentiment analysis in financial reports; (2) Application on generative tasks heavily relies on r…

Improving Child Speech Recognition and Reading Mistake Detection by Using Prompts
Lingyun Gao, Cristian Tejedor-Garcia, Catia Cucchiarini, Helmer Strik
https://arxiv.org/abs/2506.11079

Improving Child Speech Recognition and Reading Mistake Detection by Using Prompts
Automatic reading aloud evaluation can provide valuable support to teachers by enabling more efficient scoring of reading exercises. However, research on reading evaluation systems and applications remains limited. We present a novel multimodal approach that leverages audio and knowledge from text resources. In particular, we explored the potential of using Whisper and instruction-tuned large language models (LLMs) with prompts to improve transcriptions for child speech recognition, as well as …

Exactly solvable spin liquids in Kitaev bilayers and moir\'e superlattices
Ivan Dutta, Anamitra Mukherjee, Onur Erten, Kush Saha
https://arxiv.org/abs/2506.04181

Exactly solvable spin liquids in Kitaev bilayers and moiré superlattices
Building on the recent advancements on moiré superlattices, we propose an exactly solvable model with Kitaev-type interactions on a bilayer honeycomb lattice for both AA stacking and moiré superlattices. Employing Monte Carlo simulations and variational analysis, we uncover a rich variety of phases where the intra and interlayer $\mathbb{Z}_2$ fluxes (visons) are arranged in a periodic fashion in the ground state, tuned by interlayer coupling and out-of-plane external magnetic field. We furth…

This https://arxiv.org/abs/2409.18858 has been replaced.
initial toot: https://mastoxiv.page/@arXiv_csCR_…

Predicting memorization within Large Language Models fine-tuned for classification
Large Language Models have received significant attention due to their abilities to solve a wide range of complex tasks. However these models memorize a significant proportion of their training data, posing a serious threat when disclosed at inference time. To mitigate this unintended memorization, it is crucial to understand what elements are memorized and why. This area of research is largely unexplored, with most existing works providing a posteriori explanations. To address this gap, we pro…