Cross-Architecture Generalization

Follow this topic Ask about this topic

Momentum

9 papers in the last four weeks, up 125% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 32

All topics

Measuring and Reducing Cross-Vendor Mismatch in Language Models
Oct 4, 2026Erland Hilman Fuadi, Chong Tian, Xiaosong Ma +1Cross-Architecture Generalization Training-Inference Mismatch
Mohamed bin Zayed University of Artificial Intelligence
CrossGMN: Graph Metanetworks for Cross-Architecture Weight-Space Transformations
Oct 1, 2026Adir Dayan, Yam Eitan, Haggai MaronEquivariant Neural Networks Cross-Architecture Generalization
Technion – Israel Institute of Technology · NVIDIA Research
Loss-Guided Pretraining Data Selection for Time-Series Foundation Models
Sep 29, 2026Yike Li, Shaoxu Song, Jianmin WangTime Series Foundation Models Pretraining
Tsinghua University
Do Emotion Concepts Generalize Across Sources, Modalities, and Architectures in Vision-Language Models?
Sep 28, 2026Bohao Xing, Xin Liu, Kaishen Yuan +5Emotion Recognition Emotion
Lappeenranta-Lahti University of Technology LUT · Shanghai Jiao Tong University · The Hong Kong University of Science and Technology (Guangzhou) +3
Where Hallucinations Live: A Cross-Architecture Circuit in VQ-Tokenized Vision-Language Models
Sep 24, 2026Shamanthak Hegde, Xiangrui Liu, Maitreya Patel +1Recent Vision-Language Models Object Hallucination
Arizona State University · Adobe
Quantifying the Occult: A Comparative Study of Hindu and Buddhist Deities Using Machine Learning Methods
Sep 22, 2026Ankit BhattacharjeeBodhi Cross-Architecture Generalization
Department of Mechanical Engineering Indian Institute of Technology Kharagpur
Brain Metastases Segmentation for BraTS 2026 Task 1: A Multi-Architecture Comparison
Sep 21, 2026Mahdi Islam, Musarrat TabassumBrain Tumor Segmentation Cross Validation
Center for Computational & Data Sciences, Independent University, Bangladesh · Department of Electrical and Electronic Engineering, Independent University, Bangladesh · Independent Researcher
CoRA-NAS: Coarse Ranking and Anchor-Residual Refinement for Neural Architecture Search
Sep 10, 2026Yifan Yang, Zhaoyan Wang, Zheng Gao +2Neural Architecture Search Capability-Speed Trade-Offs
University of New South Wales · Korea Advanced Institute of Science and Technology
The Platonic brain bridge hypothesis: human brain networks as an architectural prior for multimodal large language models
Sep 10, 2026Pengfei Zhang, Biao Tian, Xiangang Li +1Human Brain Networks Artificial Intelligence Models
AI Thrust, Information Hub, The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China. · Alibaba Token Hub, Alibaba Group, Beijing, China.
Unifying Depth and Width Pruning for LLMs via Binary Knapsack Optimization
Aug 13, 2026Palaash Goel, Ayan Sengupta, Akshay Nambi +1Large Language Model Compression Structured Pruning
Yardi School of Artificial Intelligence, Indian Institute of Technology Delhi · Department of Electrical Engineering, Indian Institute of Technology Delhi · Microsoft Research (MSR) India
XBridge: Entity-Grounded Latent Bridge for Heterogeneous LLM Communication
Aug 12, 2026Wooseong Yang, Wei-Chieh Huang, Weizhi Zhang +3Multi-Agent Large Language Model Systems Cross-Architecture Generalization
University of Illinois Chicago · Capital One AI Foundations · Hankuk University of Foreign Studies +1
Analysis of Federated Aggregation under Model Poisoning and Backdoor Attacks: A Reconstructed Cross-Dataset and Cross-Architecture Benchmark
Aug 11, 2026Soumya Mazumdar, Vineet Kumar Rakesh, Tapas SamantaSecure Aggregation Cross-Dataset Benchmark
Department of Computer Science and Business Systems Gargi Memorial Institute of Technology Affiliated to Maulana Abul Kalam Azad University of Technology Balarampur, Mouza Beralia, Baruipur, Kolkata 700144, West Bengal, India · Computer and Informatics Group, Variable Energy Cyclotron Centre 1/AF, Bidhannagar, Kolkata 700064, West Bengal, India · Engineering Sciences, Homi Bhabha National Institute Anushaktinagar, Mumbai, Maharashtra 400094, India
Rethinking Data Efficiency in Industrial Dense Prediction: Pretraining Coherence, Not Inductive Bias, Determines ViTs Low-Data Advantage
Aug 11, 2026Haoran Sui, Yaoyuan JiaVision Transformer Frozen Backbone
ZTE Corporation, Shenzhen, Guangdong Province, P.R. China · The University of Hong Kong, Hong Kong SAR, P.R. China
Caliber: Cross-Architecture Extraction-Cost Control for Score-Returning APIs
Aug 2, 2026Chi Wang, Hanwen Wang, Yu Xia +2Model-Agnostic Defense Cross-Architecture Generalization
University of Queensland, Brisbane, Australia · Adelaide University, Adelaide, Australia · Beijing Technology and Business University, Beijing, China +1
A Cross-Architecture Audit of Direction-Based Inference-Time Defences in Vision-Language Models
Jul 30, 2026Xiangyu Yin, Tora Bodin, Rohan Menon +1Inference-Time Defense Large Language Model Jailbreaks
Chalmers University of Technology, Gothenburg, Sweden · Carl von Ossietzky University of Oldenburg, Oldenburg, Germany
Analysing Self-Harm Representations in Language Models: a Cross-Architecture Study
Jul 24, 2026Luis Espinosa-Anke, Carla Perez-AlmendrosLarge Language Model Safety Suicide Risk
Cardiff University
Trustworthiness Costs of Domain Adaptation in Small Language Models:A Cross-Architecture Empirical Study
Jul 23, 2026Ramesh B. ParamkushamLarge Language Model Adaptation Large Language Model Reliability
Independent Researcher; Adjunct Faculty, University of the Cumberlands, Williamsburg, KY, USA
A multi-architecture study of specificity refinement and false-positive mechanism analysis in prostate MRI
Jun 29, 2026Yongbo Shu, Kewen Chen, Yifeng Yuan +5Prostate Magnetic Resonance Imaging False-Negative Rate
The Second Xiangya Hospital of Central South University, Changsha, Hunan 410011, China · School of Life Sciences, Central South University, Changsha, Hunan 410013, China · Hunan Provincial Key Laboratory of Medical Information Research (Central South University), Changsha, Hunan 410011, China +1
Actionable Activation Directions for Detecting and Mitigating Emergent Misalignment Across Language Model Families
Jun 18, 2026Abdul Rafay SyedEmergent Misalignment Large Language Model Alignment
Department of Computer Science Universität des Saarlandes Saarbrücken, Germany
Cascade Classification of Dermoscopic Images of Skin Neoplasms with Controllable Sensitivity and External Clinical Validation
Jun 11, 2026Elena S. Kozachok, Sergey S. Seregin, Aleksandr V. Kozachok +2Skin Lesion Classification Melanoma
Ivannikov Institute for System Programming of the Russian Academy of Sciences (ISP RAS), Moscow, Russia · Orel Oncological Dispensary, Orel, Russia
The Cross-Architecture Substrate: A Domain-Transcendent, Calibration-Surviving Geometric Invariant of Modern Vision Encoders
Jun 5, 2026Yousef RadwanVision Encoders Cross-Architecture Generalization
KAUST
Beyond tokens: a unified framework for latent communication in LLM-based multi-agent systems
Jun 4, 2026Yingzhuo LiuMulti-Agent Large Language Model Systems Unified Framework
Pattern Selectivity is Not Task-Causal Structure: A Cross-Architecture Mechanistic Study of Composed-Task Circuits in 1B-Class Language Models
Jun 3, 2026Yongzhong XuAblation Language Modeling
VLA-Hijack: A Transferable Patch Attack against Vision-Language-Action Models via Visual Proprioception Hijacking
May 27, 2026Jiyuan Fu, Kaixun Jiang, Jingkai Jia +7Hijacking Proprioception
Fudan University · The Hong Kong Polytechnic University
Integrated and Cross-Architecture Interpretation of LLM Reasoning
May 27, 2026Leonardo Matthew Yauw, Wei-Bin Kou, Yujiu YangLLM Reasoning Strategies Interpretability
Tsinghua Shenzhen International Graduate School · Harbin Institute of Technology, Shenzhen
Capability and Robustness Cannot Both Be Free: An Information-Theoretic Bound for Vision-Language-Action Models
May 25, 2026Jianwei TaiCross-Architecture Generalization Encoders
School of Internet, Anhui University
Diagnosing Overhead in Dispatch Operations: Cross-architecture Observatory
May 20, 2026Bole Ma, Jan Eitzinger, Harald Koestler +1Load Balancing Bottlenecks
Adversarial Fragility and Language Vulnerability in Clinical AI: A Systematic Audit of Diagnostic Collapse Under Imperceptible Perturbations and Cross-Lingual Drift in Low-Resource Healthcare Settings
May 16, 2026Anthonio Oladimeji Gabriel, Ahmad Rufai YusufCross-Lingual Consistency Cross-Architecture Generalization
Centre for Clinical Intelligence & Safety, Ìyàwó, Nigeria · Tomorrow University of Applied Sciences, Germany
DIVER: Diving Deeper into Distilled Data via Expressive Semantic Recovery
May 12, 2026Qianxin Xia, Zhiyong Shu, Wenbo Jiang +3Diffusion-Based Dataset Distillation Symbolic Distillation
University of Electronic Science and Technology of China, Chengdu, China · Institute of High Performance Computing (IHPC), Agency for Science, Technology and Research (A*STAR), Singapore · Centre for Frontier AI Research (CFAR), Agency for Science, Technology and Research (A*STAR), Singapore.
Understanding Adversarial Transferability in Vision-Language Models for Autonomous Driving: A Cross-Architecture Analysis
Apr 30, 2026David Fernandez, Pedram MohajerAnsari, Amir Salarpour +1Autonomous Driving Adversarial Training
School of Computing, Clemson University, USA
Evaluating CUDA Tile for AI Workloads on Hopper and Blackwell GPUs
Apr 25, 2026Divakar Kumar Yadav, Tian Zhao, Deepak KumarCuda Graphics Processing Unit Architectures
University of Wisconsin–Milwaukee Milwaukee, WI, USA · Illinois Institute of Technology Chicago, IL, USA
Shuttling Compiler for Trapped-Ion Quantum Computers Based on Fine-Tuned Large Language Models
Dec 19, 2025Fabian Kreppel, Reza Salkhordeh, Ferdinand Schmidt-Kaler +1Quantum Processor Large Language Model Quantization
Institute of Computer Science, Johannes Gutenberg University, Staudingerweg 9, 55128 Mainz, Germany · Institute of Physics, Johannes Gutenberg University, Staudingerweg 7, 55128 Mainz, Germany · Department of Computer Science, Saarland University, Saarland Informatics Campus, E1 1, 66123 Saarbrücken, Germany

Oct 4, 2026, cs.LGJ/K move · Enter open · S save

Measuring and Reducing Cross-Vendor Mismatch in Language Models

Erland Hilman Fuadi, Chong Tian, Xiaosong Ma, Qirong Ho

Mohamed bin Zayed University of Artificial Intelligence

Cross-Architecture Generalization Training-Inference Mismatch

Open paper AskPDF arXiv

Running the same language model on different graphics processing unit (GPU) vendors can produce different logits, even when the model weights and inputs are the same. We analyze cross-vendor mismatch in two dense and two mixture-of-experts (MoE) models with five metric families, namely bitwise equality, logit differences, top-K consistency, token agreement, and task accuracy. We trace one source of the mismatch to accumulation order inside vendors' matrix instructions. Upcasting to FP32 reduces the dense model's logit error by 43% at three times the runtime, yet keeping only the MLPs in BF16 retains 94% of this gain at 1.3 times the runtime, so most of the cost of full upcasting buys little. In the MoE models, FP32 and FP16 both lower the probability error but raise the logit error and change expert selection, and FP16 fails in the dense model. An output-head low-rank adapter (LoRA) does not help either, since the final hidden state does not predict the mismatch. The mismatch also carries into training. With every seed fixed, a student distilled from a teacher running on AMD answers 431 MMLU questions differently from one distilled from the same teacher on NVIDIA. Under FP32 upcasting, bitwise equality barely changes while the output distributions move most of the way to the reference, so judging cross-vendor agreement by a single measure misreads both its cost and its gains. Code is available at https://github.com/crova-project/crova.

Momentum

Related topics

Latest papers 32