Ax Shijun Li, Wooseong Yang, Yu Wang, Tianxin Wei, Joydeep Ghosh 5/11/2026

RRCM: Ranking-Driven Retrieval over Collaborative and Meta Memories for LLM Recommendation

Proposes RRCM framework for LLM-based recommender systems using ranking-driven retrieval over collaborative and meta memories.

Ax Deeraj S K, Sadhana Devarajan, Krishna Mehra, Sudhakar Mishra 5/11/2026

Can You Break RLVER? Probing Adversarial Robustness of RL-Trained Empathetic Agents

Constructs Adversarial Empathy Benchmark to probe robustness of RL-trained empathetic language agents against adversarial user interactions.

Ax Jialin Yang, Jiankun Wang, Jiajun Wu, Henry Leung, Jiayu Zhou, Steve Drew 5/11/2026

Structural Rationale Distillation via Reasoning Space Compression

Proposes Distillation through Reasoning Path Compression to improve consistency of teacher rationales when distilling reasoning into smaller LLMs.

Ax Zixuan Xie, Xinyu Liu, Shangtong Zhang 5/11/2026

MathlibPR: Pull Request Merge-Readiness Benchmark for Formal Mathematical Libraries

Introduces MathlibPR benchmark for evaluating LLM-assisted pull request merge-readiness in Lean formal mathematics library.

Ax Junseon Yoo 5/11/2026

Topic Is Not Agenda: A Citation-Community Audit of Text Embeddings

Audits text embeddings using citation graphs of 3.58M papers, revealing disconnect between cosine similarity and conceptual relatedness in RAG.

Ax Huaiyuan Qin, Muli Yang, Gabriel James Goenawan, Peng Hu, Chen Gong, Xi Peng, Hongyuan Zhu 5/11/2026

Attention Transfer Is Not Universally Effective for Vision Transformers

Benchmarks attention transfer across 20 Vision Transformer teachers, finding it not universally effective across ViT families.

Ax Bincheng Peng, Guang Li, Ping Liu, Takahiro Ogawa, Miki Haseyama 5/11/2026

Closed-Form Linear-Probe Dataset Distillation for Pre-trained Vision Models

Develops closed-form dataset distillation method for linear probing on frozen pre-trained vision model encoders.

Ax Junichiro Niimi 5/11/2026

Three-in-One World Model: Energy-Based Consistency, Prediction, and Counterfactual Inference for Marketing Intervention

Proposes Three-in-One world model combining Deep Boltzmann Machine with task-specific heads for marketing intervention modeling.

Ax Srikar Kashyap Pulipaka 5/11/2026

PSK@EEUCA 2026: Fine-Tuning Large Language Models with Synthetic Data Augmentation for Multi-Class Toxicity Detection in Gaming Chat

System for toxicity detection in gaming chat using fine-tuned LLMs with LoRA and synthetic data augmentation across six toxicity classes.

Ax Akshita Singh, Prabesh Paudel, Siddhartha Roy 5/11/2026

Hallucination Detection via Activations of Open-Weight Proxy Analyzers

Introduces proxy-analyzer framework using open-weight models to detect hallucinations in LLM outputs via internal activations.

Ax Youngsik Yoon, Sungjae Lee, Seockbean Song, Siwei Wang, Wei Chen, Jungseul Ok 5/11/2026

PaT: Planning-after-Trial for Efficient Test-Time Code Generation

Proposes Planning-after-Trial adaptive policy for test-time code generation that efficiently allocates compute based on problem difficulty.

Ax Al Muhit Muhtadi, Mostafa Rifat Tazwar 5/11/2026

MIPIAD: Multilingual Indirect Prompt Injection Attack Defense with Qwen -- TF-IDF Hybrid and Meta-Ensemble Learning

Presents MIPIAD defense framework against multilingual indirect prompt injection attacks in RAG and tool-using LLM systems using ensemble learning.

Ax Bingqing Jiang, Difan Zou 5/11/2026

Structured Role-Aware Policy Optimization for Multimodal Reasoning

Introduces structured role-aware policy optimization for multimodal reasoning in large vision-language models using reinforcement learning from verifiable rewards.

Ax Kevin Kurian Thomas Vaidyan, Siddharth Rout 5/11/2026

Sparse Random-Feature Neural Networks with Krylov-Based SVD for Singularly Perturbed ODE

Proposes sparse random-feature neural networks with Krylov-based SVD for solving singularly perturbed ODEs with improved scalability.

Ax Mana Sakai, Masaaki Imaizumi 5/11/2026

Spectrum-Adaptive Generalization Bounds for Trained Deep Transformers

Analyzes generalization bounds for trained Transformers using spectrum-adaptive methods to improve upon existing norm-based complexity bounds.

Ax Sum Kyun Song, Bong Gyun Shin, Jae Yong Lee 5/11/2026

Discovering Ordinary Differential Equations with LLM-Based Qualitative and Quantitative Evaluation

DoLQ: method for discovering ordinary differential equations combining symbolic regression with LLM-based qualitative evaluation.

Ax Sachin Kumar 5/11/2026

Activation Differences Reveal Backdoors: A Comparison of SAE Architectures

Sparse autoencoder architectures (Crosscoders, Diff-SAE) for detecting backdoor attacks in language models via mechanistic interpretability.

Ax Sibayan Mitra (BITS Pilani), Dhruv Kumar (BITS Pilani) 5/11/2026

Mean-Pooled Cosine Similarity is Not Length-Invariant: Theory and Cross-Domain Evidence for a Length-Invariant Alternative

Analysis showing mean-pooled cosine similarity is length-dependent; proposes length-invariant alternative for comparing neural representations.

Ax Mohammad Hasan Ahmadilivani, Marten Roots, Marco Restifo, Sven-Markus Loorits, Luca Di Mauro, Jaan Raik 5/11/2026

Effective and Memory-Efficient Alternatives to ECC for Reliable Large-Scale DNNs

Memory-efficient alternatives to error correction codes for protecting deep learning models against hardware faults.

Ax Hao Wang, Yiqun Sun, Pengfei Wei, Lawrence B. Hsieh, Daisuke Kawahara 5/11/2026

Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

Sparse autoencoders as lightweight detection mechanism for adversarial attacks on vision-language models.

Ax Jinhao Jing, Qiannian Zhao, Chao Huang, Zhan Su 5/11/2026

Efficient Data Selection for Multimodal Models via Incremental Optimization Utility

Data selection framework for large multimodal models using incremental optimization utility ranking instead of LLM-as-Judge.

Ax Yubin Wu, Zicheng Cai, Liping Ning, Hua Wang, Zhi Chen, Yaohua Tang, Hao Chen 5/11/2026

LiteGUI: Distilling Compact GUI Agents with Reinforcement Learning

Reinforcement learning approach for distilling compact GUI agents that work on-device across platforms.

Ax Christopher Ries, Moussa Kassem Sbeyti, Nicolas Bianco, Nadja Klein 5/11/2026

Probabilistic Object Detection with Conformal Prediction

Conformal prediction framework for uncertainty quantification in object detection with finite-sample coverage guarantees.

Ax Nong Minh Hieu, Antoine Ledent 5/11/2026

A Refined Generalization Analysis for Extreme Multi-class Supervised Contrastive Representation Learning

Theoretical analysis of sample complexity in contrastive representation learning with dependent tuple sampling.

Ax Zhengyang Tang, Yi Zhang, Chenxin Li, Xin Lai, Pengyuan Lyu, Yiduo Guo, Weinong Wang, Junyi Li, Yang Ding, Huawen Shen, Zhengyao Fang, Xingran Zhou, Liang Wu, Fei Tang, Sunqi Fan, Shangpin Peng, Zheng Ruan, Anran Zhang, Benyou Wang, Chengquan Zhang, Han Hu 5/11/2026

Safe, or Simply Incapable? Rethinking Safety Evaluation for Phone-Use Agents

Framework for evaluating safety vs. capability in phone-use agents, addressing ambiguity in existing benchmarks.

Ax Marcel Binz, Elif Akata, Abdullah Almaatouq, Mohammed Alsobay, Oleksii Ariasov, Franziska Br\"andle, David Broska, Jason W. Burton, Nuno Busch, Frederick Callaway, Vanessa Cheung, Brian Christian, Julian Coda-Forno, Can Demircan, Vittoria Dentella, Maria K. Eckstein, No\'emi \'Eltet\H{o}, Michael Franke, Thomas L. Griffiths, Fritz G\"unther, Susanne Haridi, Sebastian Hellmann, Stefan Herytash, Linus Hof, Eleanor Holton, Isabelle Hoxha, Zak Hussain, Akshay Jagadish, Elif Kara, Valentin Kriegmair, Evelina Leivada, Li Ji-An, Tobias Ludwig, Maximilian Maier, Marcelo G. Mattar, Marvin Mathony, Alireza Modirshanechi, Robin Na, Mariia Nadverniuk, Antonios Nasioulas, Surabhi S. Nath, Helen Niemeyer, Kate Nussenbaum, Sebastian Olschewski, Thorsten Pachur, Stefano Palminteri, Aliona Petrenco, Camille V. Phaneuf-Hadd, Angelo Pirrone, Manuel Rausch, Laura Raveling, Shashank Reddy, Milena Rmus, Evan M. Russek, Tankred Saanum, Kai Sandbrink, Louis Schiekiera, Johannes A. Schubert, Luca M. Schulze Buschoff, Nishad Singhi, Leah H. Somerville, Mikhail S. Spektor, Xin Sui, Christopher Summerfield, Mirko Thalmann, Anna I. Thoma, Taisiia Tikhomirova, Vuong Truong, Polina Tsvilodub, Konstantinos Voudouris, Robert C. Wilson, Kristin Witte, Shuchen Wu, Dirk U. Wulff, Hua-Dong Xiong, Songlin Xu, Lance Ying, Xinyu Zhang, Jian-Qiao Zhu, Eric Schulz 5/11/2026

Post-training makes large language models less human-like

Study showing post-training reduces LLM alignment with human behavior; introduces Psych-201 dataset for measuring behavioral alignment at scale.

Ax Valeriy Vyaltsev, Alsu Sagirova, Anton Andreychuk, Yuri Kuratov, Konstantin Yakovlev, Aleksandr Panov, Alexey Skrynnik 5/11/2026

Learning to Communicate Locally for Large-Scale Multi-Agent Pathfinding

Decentralized multi-agent pathfinding solver using learned local communication for scalable trajectory planning in robotics and logistics.

Ax Yinsheng Yao, Jiehao Tang, Zhaozhen Yang, Dawei Cheng 5/11/2026

MAVEN: Multi-Agent Verification-Elaboration Network with In-Step Epistemic Auditing

MAVEN multi-agent verification network with epistemic auditing for reasoning tasks, enabling intermediate verification in CoT traces.

Ax Naoto Iwase, Yuki Ichihara, Mohammad Atif Quamar, Junpei Komiyama 5/11/2026

Reliable Chain-of-Thought via Prefix Consistency

Prefix consistency method for improving Chain-of-Thought reliability by using answer reproduction as verification signal.

Ax Florian A. D. Burnat, Brittany I. Davidson 5/11/2026

Quotient Semivalues for False-Name-Resistant Data Attribution

Data valuation method using quotient semivalues to resist false-name manipulation in ML data attribution.

Ax Florian A. D. Burnat 5/11/2026

Differentially Private Auditing Under Strategic Response

Game-theoretic analysis of differential privacy auditing when audited developers can strategically respond to audit queries.

Ax Yanis Merzouki, Coral Izquierdo, Matei Ignuta-Ciuncanu, Marcos Gomez-Bracamonte, Riccardo Maggioni, Alessandro Lombardi, Camilla Mazzoleni, Federico Martelli, Balazs Gunther, Jonas Petersen, Philipp Petersen 5/11/2026

FactoryBench: Evaluating Industrial Machine Understanding

FactoryBench benchmark evaluates time-series models and LLMs on industrial robotic telemetry along four causal levels.

Ax Victor Conchello Vendrell, Arnau Padres Masdemont, Niccol\`o Grillo, Jordi Ros-Giralt, Arash Behboodi, Fabio Valerio Massoli 5/11/2026

Memory-Efficient Looped Transformer: Decoupling Compute from Memory in Looped Language Models

Memory-efficient recurrent LLM architecture decoupling compute from memory in looped reasoning models with sublinear KV cache.

Ax Sree Bhattacharyya, Samarth Khanna, Leona Chen, Lucas Craig, Tharun Dilliraj, James Z. Wang 5/11/2026

Beyond Confidence: Rethinking Self-Assessments for Performance Prediction in LLMs

Research on LLM self-assessment using cognitive appraisal theory as alternative to confidence scores for performance prediction.

Ax Dimitrios Mallis, Marco Wang, Ahmet Serdar Karadeniz, Elisa Ricci, Anis Kacem, Djamila Aouada 5/11/2026

Text-to-CAD Evaluation with CADTests

CADTestBench introduces first test-based evaluation benchmark for Text-to-CAD task using automated testing methodology.

Ax Ionut-Vlad Modoranu, Mher Safaryan, Dan Alistarh 5/11/2026

MatryoshkaLoRA: Learning Accurate Hierarchical Low-Rank Representations for LLM Fine-Tuning

arXiv: MatryoshkaLoRA - adaptive rank LoRA for efficient LLM fine-tuning without grid search. Parameter-efficient training.

Ax Kandula Eswara Sai Kumar, Parth Dhananjay Danve, Abhishek Chopra, Rut Lineswala 5/11/2026

Exploring the non-convexity in machine learning using quantum-inspired optimization

arXiv: Quantum-inspired optimization for non-convex ML problems in high dimensions. Theoretical optimization approach.

Ax Zekun Wu (University College London), Ze Wang (University College London), Seonglae Cho (Holistic AI), Yufei Yang (Imperial College London), Adriano Koshiyama (University College London), Sahan Bulathwela (University College London), Maria Perez-Ortiz (University College London) 5/11/2026

Tool Calling is Linearly Readable and Steerable in Language Models

arXiv research: Tool-calling in LLMs is linearly readable/steerable via internal activations. Tested 12 models, 77-100% steering accuracy.

Ax Yiyang Shen, Yutian He, Weiran Wang, Qihang Lin 5/11/2026

Penalty-Based First-Order Methods for Bilevel Optimization with Minimax and Constrained Lower-Level Problems

First-order optimization methods for bilevel problems with minimax structures in upper and lower levels.

Ax Himanshu Udupi, Xiaocong Yang, ChengXiang Zhai 5/11/2026

Globally Optimal Training of Spiking Neural Networks via Parameter Reconstruction

Global training method for Spiking Neural Networks via parameter reconstruction addressing surrogate gradient approximation errors.

Ax Ying Shen, Tianrong Chen, Yuan Gao, Yizhe Zhang, Yuyang Wang, Miguel \'Angel Bautista, Shuangfei Zhai, Joshua M. Susskind, Jiatao Gu 5/11/2026

STARFlow2: Bridging Language Models and Normalizing Flows for Unified Multimodal Generation

STARFlow2 unifying language models and normalizing flows for multimodal text-image generation with autoregressive normalizing flows.

Ax R\"uveyda Yilmaz, Yuli Wu, Johannes Stegmaier, Volkmar Schulz 5/11/2026

PET-Adapter: Test-Time Domain Adaptation for Full and Limited-Angle PET Image Reconstruction

PET-Adapter framework for test-time domain adaptation in medical image reconstruction handling Poisson noise and limited-angle acquisitions.

Ax Houssam Zenati, Arthur Gretton 5/11/2026

Semiparametric Efficient Test for Interpretable Distributional Treatment Effects

DR-ME test for interpretable distributional treatment effects, first semiparametrically efficient finite-location test for detecting distributional shifts.

Ax Julie Kallini, Artidoro Pagnoni, Tomasz Limisiewicz, Gargi Ghosh, Luke Zettlemoyer, Christopher Potts, Xiaochuang Han, Srinivasan Iyer 5/11/2026

Fast Byte Latent Transformer

Byte Latent Transformer (BLT) addressing slow byte-level autoregressive generation with diffusion-based training and generation techniques.

Ax Jane H. Lee, Anay Mehrotra, Manolis Zampetakis 5/11/2026

A Note on Non-Negative $L_1$-Approximating Polynomials

Mathematical study of non-negative L1-approximating polynomials under Gaussian distributions with applications to computational learning theory.

Ax Jiatao Gu, Tianrong Chen, Ying Shen, David Berthelot, Shuangfei Zhai, Josh Susskind 5/11/2026

Normalizing Trajectory Models

Normalizing Trajectory Models (NTM) for efficient diffusion-based generation with few steps while preserving likelihood framework.

Ax Bhavesh Kalisetti, Vincent Wang, Gaurav R. Ghosal, Maryam Bijanzadeh, Reza Abbasi-Asl 5/11/2026

Multi-Stage Prototype Learning for Interpretable Time Series Classification

Multi-stage prototype learning framework for interpretable multivariate time series classification identifying predictive temporal patterns.

Ax Hongyuan Zhang, Yanchen Xu, Sida Huang, Xuelong Li 5/11/2026

Data Augmentation of Contrastive Learning is Estimating Positive-incentive Noise

Theoretical analysis connecting contrastive learning data augmentation to positive-incentive noise estimation via information theory.

Ax Zhichao Wang, Bin Bi, Can Huang, Shiva Kumar Pentyala, Zixu James Zhu, Sitaram Asur, Na Claire Cheng, Cheng Wan, Dong Nie, Lingzi Hong 5/11/2026

UNA: A Unified Supervised Framework for Efficient LLM Alignment Across Feedback Types

UNA framework unifying diverse feedback types (preferences, scores, scalars) for efficient LLM alignment across RLHF and DPO methods.

Ax Surbhi Goel, Adam R. Klivans, Konstantinos Stavropoulos, Arsen Vasilyan 5/11/2026

Testing Noise Assumptions of Learning Algorithms

Algorithm for testing whether training data satisfies noise model assumptions in computational learning theory.