The paper introduces open-weight language models ranging from 0.6B to 14B parameters, specifically adapted for Wolof—a language spoken by 23 million people in Senegal—through instruction tuning, preference alignment, and model merging. The authors detail the development pipeline and dataset, and report extensive evaluations on 63 public and in-house tasks, demonstrating strong Wolof understanding and generation while maintaining performance on other languages and cross‑lingual tasks. The 8B and 14B models are noted to be competitive with much larger commercial systems, and the models are currently being tested in internal customer‑care pilots and are publicly available on Hugging Face.
By Gw\'enol\'e Lecorv\'e, Delphine Charlet, Pierre Adam, Floriane Behanzin, Lionel Delphin-Poulat, Johannes Heinecke, G\'eraldine Damnati, Pelin Kaytmaz, Nourredine Khaous, Charbel Kindji, Meriam Maadi, Abdoulaye Mbaye, Claire Perroux, Isma\"el Rousseau, Papa S\'ega Wad\'e, Gilles Xiberras, Olivier Zhang, Malek Zitouni
Multi‑SFT is a new, commercially permissive dataset comprising 502,818 instruction‑response pairs in five European languages, created through translation and regeneration pipelines to address gaps where simple translation is insufficient. Human evaluation shows strong inter‑annotator agreement (≥0.87) across all languages. Fine‑tuning nine models from three families on Multi‑SFT yields performance competitive with larger, proprietary mixes and official instruction‑tuned releases, especially on safety‑related benchmarks.
By Daniela Baiamonte, Kamyar Zeinalipour, Elena Fano, Stefano Simonazzi, Andrea Zugarini
The study investigates how different rating scale designs affect scores produced by large language models (LLMs) in text scoring tasks. By varying format, range, and sign across tasks like political leaning, sentiment, and toxicity, the authors find that a single design change can shift normalized scores by about 0.10 and alter the category of roughly 20% of texts. The research highlights that rubric-based prompting does not mitigate this sensitivity and emphasizes the need for robustness checks when using LLMs for scalar scoring.
By Wonjeong Jo, Gyuwan Kim, Ming Shan Hee
PathCal is a training‑free controller that refines large reasoning language models by dynamically adjusting the influence of reflection markers—continuation, revision, and alternative‑opening—during inference. By applying finite logit adjustments without extra forward passes, it exploits instantaneous competition among these markers to improve the accuracy–length tradeoff in mathematical reasoning tasks and shorten generation in coding and multilingual math. The method is grounded in an exact log‑odds identity and frozen‑prefix probes that confirm its local directional effect.
By Lingyu Jiang, Zirui Li, Lingyu Xu, Shuo Xing, Peiran Li, Tsubasa Takahashi, Dengzhe Hou, Zhengzhong Tu, Kazunori Yamada, Fangzhou Lin
ZeST is a zero‑shot traversability navigation system that uses a vision‑language model (VLM) to generate stochastic measurements, which are fused into an uncertainty‑aware posterior. By avoiding real‑world data collection, ZeST offers a cost‑effective, scalable solution for predicting terrain traversability. Experiments in indoor and outdoor settings show ZeST achieving 90–100% success rates, outperforming baselines that top 60%.
By Shreya Gummadi, Mateus V. Gasparino, Gianluca Capezzuto, Marcelo Becker, Girish Chowdhary
The paper introduces Grounding IDs, latent identifiers that emerge when external visual cues—such as symbols or grid lines—structure both visual and textual modalities in large vision-language models. These IDs improve the alignment between visual components and their textual descriptions, enhancing cross‑modal grounding, reducing hallucinations, and boosting overall performance on multimodal tasks. The study demonstrates that partition‑inducing cues lead to more accurate connections and better attention between related components.
By Amirmohammad Izadi, Hosein Hasani, Fatemeh Askari, Mobin Bagherian, Sadegh Mohammadian, Mohammad Izadi, Mahdieh Soleymani Baghshah
The article reviews how cardiac imaging techniques such as echocardiography, CMR, and CT can be integrated with bulk, single-cell, and spatial transcriptomics to create joint representations that link imaging phenotypes to detailed molecular states. It outlines the representation needs of each modality, compares multimodal fusion strategies, and synthesizes pipelines for radiogenomics, spatial alignment, and image-based gene‑expression prediction, while discussing validation, limitations, and failure modes. The review highlights emerging spatial multiomic maps of human myocardium and atherosclerotic plaque, noting that cost, scalability, and tissue availability remain major barriers to large‑scale cardiovascular translation.
By Minh H. N. Le, Thanh-Huy Nguyen, Tao Li, Bao Quang Gia Le, Han H. Huynh, Monika Raj, Carl Yang, Min Xu, Tuan Vinh, Nguyen Quoc Khanh Le
ChaosBench-Logic v2 is a 40,886-question benchmark designed to evaluate large language models’ logical reasoning about dynamical systems, featuring 165 systems, 27 first‑order logic predicates, and 78 axiom edges. The accompanying CARE protocol highlights failure modes such as prior collapse, paraphrase inconsistency, and difficulty with parameter‑dependent dynamics. In tests of 14 models, regime‑transition reasoning remains near random (MCC = 0.05), while first‑order deduction with given premises achieves MCC = 0.52, with proprietary models excelling in cross‑indicator and consistency tasks and open‑source Qwen 2.5‑32B leading indicator diagnostics.
By Noel Thomas
The paper introduces a test‑time adaptation (TTA) method for AI text detection that addresses continual distribution shifts such as adversarial humanization, new large language models, and temporal drift in human writing. By exploiting inference‑time homogeneity among unlabeled samples, the TTA approach achieves robust performance, outperforming state‑of‑the‑art supervised detectors—e.g., improving detection of adversarial AI‑generated text from 27.0% to 93.3%. The study demonstrates that TTA can maintain high detection accuracy in real‑world, evolving deployment scenarios.
By Kevin Ren, Manish Raghavan, Nikhil Garg
arXiv:2610. 21312v1 Announce Type: new Abstract: Data-free preconditioning builds the Kronecker factors of KFAC from synthetic probes instead of private data.
By Shilin Zhang, Kangning Wang
P2TPO introduces a structure‑aligned credit‑assignment framework for full‑page document parsing, aligning block‑level quality contrasts with generated spans and redistributing credit within matched blocks. The method employs Unified Anchor Pricing to handle matched and missing elements and Span Credit Redistribution to adjust token‑level credit while preserving overall block advantage. Experiments show P2TPO surpasses existing policy optimization baselines across six parsing benchmarks and scales well with model size and initialization.
By Yong Zu, Ninghua Yang, Zezhong Tan, Xinhong Ma, Ziqiang Dong, Yang Yang
The paper introduces the Single-Increment Rejection Sampler (SIRS) to improve query efficiency in parallel masked diffusion models (MDMs). SIRS estimates the log acceptance probability using only one randomly sampled increment, then applies a Bernoulli factory to decide accept/reject outcomes. By Rao‑Blackwellizing each increment and bounding the KL‑divergence, SIRS reduces bias while maintaining parallel decoding performance.
By Shiyuan Zhang, Xuheng Li, Qiwei Di, Quanquan Gu
The study evaluated whether imagined speech decoded from scalp EEG can transfer across individuals using three public datasets. Across 482 participant pairs, cross‑subject decoding accuracy was essentially at chance, while within‑subject decoding only surpassed chance in one dataset. The results suggest that imagined‑speech content does not generalize between people and that larger single‑condition datasets are more beneficial than further model development.
By Blessing Isoyiza Adeika, Onyema Osuagwu
The paper introduces a general framework for non‑interactive locally differentially private (LDP) learning of structured univariate distributions that can be approximated by piece‑wise polynomials. It achieves the first sample‑efficient LDP learners for several distribution classes, notably mixtures of log‑concave distributions such as Gaussian mixture models (GMMs), while providing agnostic guarantees and computational efficiency. The authors also establish a new lower bound for learning GMMs under LDP, surpassing previous pure‑DP packing bounds, and develop an efficient non‑interactive LDP algorithm for synthetic data generation over large domains, which is then combined with a deterministic adaptation of a non‑private piecewise polynomial learner.
By Mohammad Afzali, Hassan Ashtiani, Sivan Sabato
ContextLift is a counterfactual augmentation technique that mitigates conditional emergent misalignment in language models. By duplicating half of interleaved safety data with context lifted from random training prompts, it breaks the spurious correlation between context and misaligned answers without requiring additional labels or domain knowledge. Experiments on Qwen2.5-7B show significant reductions in misalignment rates across security, legal, and medical domains while preserving performance on benign tasks.
By Florian Mai, David Kacz\'er
The paper introduces FPCD, a framework that stabilizes the modality gap in vision‑language models during continual learning. It combines Fast Null‑space Projection, which projects gradients onto approximate null spaces to reduce interference, with Contrastive Distillation that aligns image‑to‑text and text‑to‑image similarity distributions using a small reference set. On the 11‑task MTIL benchmark, FPCD achieves strong performance, maintaining stable modality gaps and superior zero‑shot retrieval on COCO Captions and Flickr30k.
By Tiantian Peng, Yuyang Liu, Shuo Yang, Qiuhe Hong, YongHong Tian
SEVerA is a three‑stage framework that synthesizes self‑evolving agents with formal safety guarantees. It introduces Formally Guarded Generative Models (FGGM) that wrap parametric models in verified rejection samplers, ensuring every output satisfies a first‑order logic contract. The framework searches for candidate programs, verifies them against hard constraints, and then fine‑tunes them to improve task utility while preserving correctness, achieving zero constraint violations on tasks such as Dafny verification, symbolic math synthesis, and policy‑compliant agentic tool use.
By Debangshu Banerjee, Changming Xu, Eugene Ie, Ming Zhang, Daiyi Peng, Chu-Cheng Lin, Tarun Suresh, Gagandeep Singh
The paper proposes a method for releasing Gaussian Mixture Models (GMMs) under differential privacy (DP) by minimizing Kullback‑Leibler (KL) divergence between the released and non‑private models. It jointly optimizes privacy‑budget allocation and parameter perturbations for mixture weights, means, and covariances, providing a tractable expression for KL divergence and a theoretical analysis of DP guarantees. Experiments on synthetic and real data show that the approach achieves strong privacy while preserving high utility.
By Hang Liu, Anna Scaglione, Sean Peisert
A Greedy PDE Router for Blending Neural Operators and Classical Methods proposes an approximate greedy router that selects solvers from an ensemble to solve PDEs efficiently. The method addresses the challenge of lacking true error knowledge by mimicking greedy selection, aiming to combine the strengths of machine learning and classical solvers. Experiments on Poisson, convection-diffusion, and anisotropic diffusion equations demonstrate reduced time-to-solution compared to single-solver baselines and existing hybrid approaches such as HINTS.
By Sahana Rayan, Yash Patel, Ambuj Tewari
VALCON introduces a role-driven framework to benchmark and improve value consistency in large language models (LLMs) by evaluating how well an LLM’s generative outputs align with its evaluative judgments when the same model alternates between generator and evaluator roles. The framework creates three benchmark suites—VAB-Consistency, VAB-Boundary, and VAB-Challenge—using dual-role variation to expose hidden value discrepancies. Experiments across 28 LLMs reveal widespread cross-role inconsistency, and mechanistic probing identifies a role-dependent representational decoupling; a representation-directed intervention is then proposed to mitigate these failures, establishing value consistency as a distinct alignment dimension.
By Junru Wu, Jingting Zheng, Yuqi Ren, Deyi Xiong