Ax Zhenyu Zhao, Aparna Balagopalan, Adi Agrawal, Dilshoda Yergasheva, Waseem Alshikh, Daniel M. Bikel 4/30/2026

The Price of Agreement: Measuring LLM Sycophancy in Agentic Financial Applications

Evaluation of sycophancy in LLM-based financial agents, measuring tendency to agree with users over correctness in agentic tasks.

Ax Junxing Hu, Tianlong Li, Lei Yu, Ai Han 4/30/2026

OxyGent: Making Multi-Agent Systems Modular, Observable, and Evolvable via Oxy Abstraction

Open-source framework for building modular, observable, and evolvable multi-agent systems with pluggable components and evolution engine.

Ax Wenxuan Wang, Yuk-Kit Chan, Zixuan Ling, Juluan Shi, Youliang Yuan, Jen-tse Huang, Yifei Zhang, Wenxiang Jiao, Zhaopeng Tu, Michael R. Lyu 4/30/2026

Identifying the Achilles' Heel: An Iterative Method for Dynamically Uncovering Factual Errors in Large Language Models

Iterative method for dynamically identifying and uncovering factual errors in large language models without extensive human labeling.

Ax Yunkun Zhang, Jin Gao, Zheling Tan, Lingfeng Zhou, Kexin Ding, Mu Zhou, Shaoting Zhang, Dequan Wang 4/30/2026

Data-Centric Foundation Models in Computational Healthcare: A Survey

Survey of data-centric foundation models in computational healthcare, emphasizing data quality and characterization.

Ax Wenxuan Wang, Juluan Shi, Zixuan Ling, Yuk-Kit Chan, Chaozheng Wang, Cheryl Lee, Youliang Yuan, Jen-tse Huang, Wenxiang Jiao, Michael R. Lyu 4/30/2026

Learning to Ask: When LLM Agents Meet Unclear Instruction

Framework for LLM agents to request clarification when faced with ambiguous or incomplete instructions before executing tools.

Ax Ningyuan Xi, Yetao Wu, Kun Fan, Teng Chen, Qingqing Gu, Luo Ji 4/30/2026

A Practice of Post-Training on Llama-3 70B with Optimal Selection of Additional Language Mixture Ratio

Study on optimal language mixture ratios for continual pre-training of Llama-3 70B across multiple languages and domains.

Ax Yuxia Wang, Rui Xing, Jonibek Mansurov, Giovanni Puccetti, Zhuohan Xie, Minh Ngoc Ta, Jiahui Geng, Jinyan Su, Mervat Abassy, Saad El Dine Ahmed, Kareem Elozeiri, Nurkhan Laiyk, Maiya Goloburda, Tarek Mahmoud, Raj Vardhan Tomar, Alexander Aziz, Ryuto Koike, Masahiro Kaneko, Artem Shelmanov, Ekaterina Artemova, Vladislav Mikhailov, Akim Tsvigun, Alham Fikri Aji, Nizar Habash, Iryna Gurevych, Preslav Nakov 4/30/2026

Is Human-Like Text Liked by Humans? Multilingual Human Detection and Preference Against AI

Multilingual study evaluating human ability to detect LLM-generated text across 16 datasets covering 9 languages and domains.

Ax Lin Sun, Guangxiang Zhao, Xiaoqi Jian, Yuhan Wu, Weihong Lin, Yongfu Zhu, Qilong Shi, Change Jia, Aomufei Yuan, Yuxuan Tian, Linglin Zhang, Jinzhu Wu, Junfeng Ran, Sai-er Hu, Zihan Jiang, Junting Zhou, Wenrui Liu, Xusen Xiao, Bin Cui, Tong Yang, Xiangzheng Zhang 4/30/2026

TinyR1-32B-Preview: Boosting Accuracy with Branch-Merge Distillation

Branch-Merge distillation method for compressing large language models while maintaining performance through knowledge transfer.

Ax Christian Schroeder de Witt, Klaudia Krawiecka, Igor Krawczuk, Ben Hagag, William L. Anderson, Peter Belcak, Ben Bucknall, Xiaohong Cai, Ayush Chopra, Doron Cohen, Ron F. Del Rosario, Andis Draguns, Annie Gray, Keren Katz, Vasilios Mavroudis, Jaron Mink, Sumeet Ramesh Motwani, Jonathan Petit, Leif-Sebastian Rembeck, Chandler Smith, John Sotiropoulos, Steven Young, Sarah Scheffler, Mary Llewellyn 4/30/2026

Open Challenges in Multi-Agent Security: Towards Secure Systems of Interacting AI Agents

Analysis of security challenges in multi-agent AI systems, including collusion, swarm attacks, and privacy breaches in networked agent deployments.

Ax Ada Chen, Yongjiang Wu, Junyuan Zhang, Jingyu Xiao, Shu Yang, Jen-tse Huang, Kun Wang, Wenxuan Wang, Shuai Wang 4/30/2026

A Survey on the Safety and Security Threats of Computer-Using Agents: JARVIS or Ultron?

Survey of safety and security threats in computer-using agents, covering vulnerabilities in LLM-based systems that interact with GUIs and applications.

Ax Behnam Yousefimehr, Mehdi Ghatee, Javad Fazli, Shervin Ghaffari, Zahra Rafei, Mohammad Amin Seifi, Sajed Tavakoli, Abolfazl Nikahd, Mahdi Razi Gandomani, Alireza Orouji, Ramtin Mahmoudi Kashani, Sarina Heshmati, Negin Sadat Mousavi 4/30/2026

Data Balancing Strategies: A Systematic Survey of Resampling and Augmentation Methods

Systematic survey of data balancing methods including oversampling and augmentation techniques for addressing class imbalance in machine learning.

Ax Jong Hak Moon, Geon Choi, Paloma Rabaey, Min Gwan Kim, Jung-Oh Lee, Hyuk Gi Hong, Eun Woo Doe, Hangyul Yoon, Jiyoun Kim, Harshita Sharma, Daniel C. Castro, Javier Alvarez-Valle, Edward Choi 4/30/2026

Lunguage: A Benchmark for Structured and Sequential Chest X-ray Interpretation

arXiv paper presenting LUNGUAGE benchmark for structured chest X-ray report generation with longitudinal evaluation support.

Ax Ujjwal Upadhyay, Mukul Ranjan, Zhiqiang Shen, Mohamed Elhoseiny 4/30/2026

Time Blindness: Why Video-Language Models Can't See What Humans Can?

arXiv paper introducing SpookyBench showing video-language models struggle with temporal-only patterns when spatial information is absent.

Ax Junzhe Zhang, Huixuan Zhang, Xinyu Hu, Li Lin, Mingqi Gao, Shi Qiu, Xiaojun Wan 4/30/2026

MINOS: A Multimodal Evaluation Model for Bidirectional Generation Between Image and Text

arXiv paper on MINOS, a multimodal evaluation model for bidirectional image-text generation assessment.

Ax Samuel J. Weisenthal 4/30/2026

Treatment, evidence, imitation, and chat

arXiv paper investigating LLM potential for medical decision-making, comparing treatment problem approaches with evidence-based medicine.

Ax Mohammad Naeimi, Mostafa Haghir Chehreghani 4/30/2026

PBiLoss: Popularity-Aware Regularization to Improve Fairness in Graph-Based Recommender Systems

arXiv paper on addressing popularity bias in graph neural network recommender systems using regularized fairness approach.

Ax Wenting Chen, Guo Yu, Yiu-Fai Cheung, Meidan Ding, Jie Liu, Zizhan Ma, Wenxuan Wang, Linlin Shen 4/30/2026

Beyond the Leaderboard: Rethinking Medical Benchmarks for Large Language Models

arXiv paper introducing MedCheck, a lifecycle-oriented benchmark framework for evaluating LLMs in healthcare with clinical fidelity.

Ax Jian Xu, Yican Liu, Delu Zeng, John Paisley, Qibin Zhao 4/30/2026

Neural Bridge Processes

arXiv paper proposing Neural Bridge Processes improving expressivity of neural diffusion models with input-conditioned noise.

Ax Rui Su, Honghao Dong, Haojie Jin, Yisong Chen, Guoping Wang, Sheng Li 4/30/2026

Vertex Features for Neural Global Illumination

arXiv paper on neural vertex features for efficient learnable 3D scene representation in neural rendering.

Ax Fengyi Wu, Yifei Dong, Yilong Dai, Guangyu Chen, Qifeng Wu, Huiting Huang, Hang Wang, Qi Dai, Alexander G. Hauptmann, Zhi-Qi Cheng 4/30/2026

GoViG: Goal-Conditioned Visual Navigation Instruction Generation via Multimodal Reasoning

arXiv paper on generating visual navigation instructions from egocentric observations using multimodal reasoning.

Ax Emmanouil Kritharakis, Dusan Jakovetic, Antonios Makris, Konstantinos Tserpes 4/30/2026

Robust Federated Learning under Adversarial Attacks via Loss-Based Client Clustering

arXiv paper on federated learning robustness against Byzantine adversarial attacks with loss-based client clustering.

Ax Zoe Kotti, Konstantina Dritsa, Diomidis Spinellis, Panos Louridas 4/30/2026

The Fools are Certain; the Wise are Doubtful: Exploring LLM Confidence in Code Completion

Analysis of confidence calibration in code-completion LLMs. Evaluates relationship between model confidence and prediction accuracy for code generation.

Ax Hasibur Rahman, Smit Desai 4/30/2026

Vibe Check: Understanding the Effects of LLM-Based Conversational Agents' Personality and Alignment on User Perceptions in Goal-Oriented Tasks

Empirical study (N=150) examining how LLM conversational agent personality and alignment affect user perceptions in goal-oriented tasks like travel planning.

Ax Sigmund Hennum H{\o}eg, Aksel Vaaler, Chaoqi Liu, Olav Egeland, Yilun Du 4/30/2026

Hybrid Diffusion for Simultaneous Symbolic and Continuous Planning

Hybrid Diffusion framework combining symbolic and continuous planning for long-horizon robotic tasks using diffusion models.

Ax Younes Hourri, Mohammad Mozaffari, Maryam Mehri Dehnavi 4/30/2026

PATCH: Learnable Tile-level Hybrid Sparsity for LLMs

PATCH: Learnable sparsity method for LLM deployment combining unstructured and semi-structured pruning to reduce memory/compute costs while maintaining GPU acceleration.

Ax William Walden, Marc Mason, Orion Weller, Laura Dietz, John Conroy, Neil Molino, Hannah Recknor, Bryan Li, Gabrielle Kaili-May Liu, Yu Hou, Dawn Lawrie, James Mayfield, Eugene Yang 4/30/2026

Auto-ARGUE: LLM-Based Report Generation Evaluation

Auto-ARGUE: LLM-based evaluation tool for citation-backed report generation from RAG systems. Open-source implementation of ARGUE framework.

Ax Christoph Riedl 4/30/2026

Emergent Coordination in Multi-Agent Language Models

Information-theoretic framework to detect higher-order structure and emergent coordination in multi-agent LLM systems using data-driven analysis.

Ax Congmin Zheng, Jiachen Zhu, Zhuoying Ou, Yuxiang Chen, Kangning Zhang, Rong Shan, Zeyu Zheng, Mengyue Yang, Jianghao Lin, Yong Yu, Weinan Zhang 4/30/2026

A Survey of Process Reward Models: From Outcome Signals to Process Supervisions for Large Language Models

Survey of Process Reward Models that evaluate LLM reasoning step-by-step rather than just final outputs. Covers data generation, model building, and test-time applications.

Ax Zhezheng Hao, Hong Wang, Haoyang Liu, Jian Luo, Jiarui Yu, Hande Dong, Qiang Lin, Can Wang, Jiawei Chen 4/30/2026

Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective

Analyzes entropy collapse in reinforcement learning with verifiable rewards for LLM reasoning via entropy change perspective.

Ax Kangkang Sun, Jun Wu, Minyi Guo, Jianhua Li, Jianwei Huang 4/30/2026

FedPF: Accurate Target Privacy Preserving Federated Learning Balancing Fairness and Utility

FedPF algorithm for federated learning balancing fairness and privacy using differential privacy and zero-sum game formulation.

Ax Junwei Liu, Chen Xu, Chong Wang, Tong Bai, Weitong Chen, Kaseng Wong, Yiling Lou, Xin Peng 4/30/2026

EvoDev: An Iterative Feature-Driven Framework for End-to-End Software Development with LLM-based Agents

EvoDev iterative framework for end-to-end software development using LLM agents inspired by feature-driven development practices.

Ax Zain Muhammad Mujahid, Dustin Wright, Isabelle Augenstein 4/30/2026

Stress Testing Factual Consistency Metrics for Long-Document Summarization

Systematic evaluation of reference-free factuality metrics for long-document abstractive summarization using LLMs.

Ax Inferix Team, Tianyu Feng, Yizeng Han, Jiahao He, Yuanyu He, Xi Lin, Teng Liu, Hanfeng Lu, Jiasheng Tang, Wei Wang, Zhiyuan Wang, Jichao Wu, Mingyang Yang, Yinghao Yu, Zeyu Zhang, Bohan Zhuang 4/30/2026

Inferix: A Block-Diffusion based Next-Generation Inference Engine for World Simulation

Inferix block-diffusion inference engine for world models generating long, realistic videos for agentic and embodied AI.

Ax Hanwen Wan, Zexin Lin, Yixuan Deng, Xiaoqiang Ji 4/30/2026

Value-Guided Iterative Refinement and the DIQ-H Benchmark for Evaluating VLM Robustness

DIQ-H benchmark and value-guided refinement method for evaluating VLM robustness in adversarial conditions and continuous deployment.

Ax Shengkun Cui, Rahul Krishna, Saurabh Jha, Ravishankar K. Iyer 4/30/2026

PRAXIS: Integrating Program Analysis with Observability for Root-Cause Analysis

PRAXIS orchestrates LLM-driven agentic workflow for root-cause analysis in cloud incidents using program analysis.

Ax Sasha Ronaghi, Chloe Stanwyck, Asad Aali, Amir Ronaghi, Miguel Fuentes, Tina Hernandez-Boussard, Emily Alsentzer 4/30/2026

Training-Free Adaptation of New-Generation LLMs using Legacy Clinical Models

CAPT enables training-free domain adaptation of new-generation LLMs using legacy clinical models via model ensembling.

Ax Iago Alves Brito, Walcy Santos Rezende Rios, Julia Soares Dollis, Diogo Fernandes Costa Silva, Arlindo Rodrigues Galv\~ao Filho 4/30/2026

Safety Is Not Universal: The Selective Safety Trap in LLM Alignment

Exposes selective safety trap showing LLM safety evaluations mask vulnerabilities in underrepresented communities.

Ax Quang-Hung Bui, Anh Son Ta 4/30/2026

AdaFRUGAL: Adaptive Memory-Efficient Training with Dynamic Control

AdaFRUGAL automates memory-efficient LLM training by dynamically controlling gradient splitting hyperparameters during optimization.

Ax Hongbo Bai, Yujin Zhou, Yile Wu, Chi-Min Chan, Pengcheng Wen, Kunhao Pan, Sirui Han, Yike Guo 4/30/2026

Glance-or-Gaze: Incentivizing LMMs to Adaptively Focus Search via Reinforcement Learning

Glance-or-Gaze uses RL to enable large multimodal models to adaptively focus on relevant image regions for search-augmented queries.

Ax Chengyu Du, Xintao Wang, Aili Chen, Weiyuan Li, Rui Xu, Junteng Liu, Zishan Huang, Rong Tian, Zijun Sun, Yuhao Li, Liheng Feng, Deming Ding, Pengyu Zhao, Yanghua Xiao 4/30/2026

HER: Human-like Reasoning and Reinforcement Learning for LLM Role-playing

HER framework enables LLMs to simulate personas with human-like reasoning and reinforcement learning for role-playing tasks.

Ax Xinyue Li, Zhiming Xu, Min Tang, Zhaolin Cai, Sijing Wu, Xiongkuo Min, Yitong Chen, Guangtao Zhai 4/30/2026

ELIQ: A Label-Free Framework for Quality Assessment of Evolving AI-Generated Images

ELIQ label-free framework for assessing quality of AI-generated images without requiring manual labels.

Ax Chenghui Zou, Ning Wang, Tiesunlong Shen, Luwei Xiao, Chuan Ma, Xiangpeng Li, Rui Mao, Erik Cambria 4/30/2026

Affective Flow Language Model for Emotional Support Conversation

AFlow framework for emotional support conversations using LLMs with intermediate strategy supervision via affective flow modeling.

Ax Junbo Jacob Lian, Yujun Sun, Huiling Chen, Chaoyu Zhang, Hanzhang Qin, Chung-Piaw Teo 4/30/2026

ReLoop: Structured Modeling and Behavioral Verification for Reliable LLM-Based Optimization

ReLoop framework addresses LLM-to-code translation failures in optimization problems through structured generation and behavioral verification.

Ax Haokun Liu, Zhaoqi Ma, Yicheng Chen, Masaki Kitagawa, Wentao Zhang, Zicen Xiong, Jinjie Li, Moju Zhao 4/30/2026

CoFL: Continuous Flow Fields for Language-Conditioned Navigation

End-to-end policy using continuous flow fields for language-conditioned robot navigation in complex scenes.

Ax Toms Bergmanis, Martins Kronis, Ingus J\=anis Pretkalni\c{n}\v{s}, D\=avis Nicmanis, Je\c{l}izaveta Jelinska, Roberts Rozis, Rinalds V\=iksna, M\=arcis Pinnis 4/30/2026

TildeOpen LLM: Leveraging Curriculum Learning to Achieve Equitable Language Representation

TildeOpen LLM: 30B-parameter open-weight model trained on 34 European languages using curriculum learning for equitable linguistic representation.

Ax Zhen Zhang, Jielei Chu, Jiangtao Hu, Bin Liu, Jie Wang, Ya Liu, Tianrui Li 4/30/2026

Causally Sufficient and Necessary Feature Expansion for Class-Incremental Learning

Addresses feature collisions in class-incremental learning using causal analysis to identify spurious correlations and expand features.

Ax Maxwell Miller-Golub, Collin Coil, Kamil Faber, Marcin Pietron, Panpan Zheng, Pasquale Minervini, Roberto Corizzo 4/30/2026

Rethinking the Harmonic Loss via Non-Euclidean Distance Layers

Proposes non-Euclidean distance layers to improve harmonic loss function beyond cross-entropy for deep neural network training.

Ax Ziyu Chen, Yilun Zhao, Chengye Wang, Rilyn Han, Manasi Patwardhan, Arman Cohan 4/30/2026

SciMDR: Advancing Scientific Multimodal Document Reasoning

SciMDR introduces synthesize-and-reground framework for constructing scientific multimodal document reasoning datasets balancing scale, faithfulness, and realism.

Ax Chenlu Ye, Xuanchang Zhang, Yifan Hao, Zhou Yu, Ziji Zhang, Abhinav Gullapalli, Hao Chen, Jing Huang, Tong Zhang 4/30/2026

Adaptive Layerwise Perturbation: Unifying Off-Policy Corrections for LLM RL

Adaptive Layerwise Perturbation unifies off-policy corrections for LLM reinforcement learning to address policy staleness and training-inference mismatch.

Ax Ga\"etan Hadjeres, Marc Ferras, Khaled Koutini, Benno Weck, Alexandre Bittar, Thomas Hummel, Zineb Lahrichi, Hakim Missoum, Joan Serr\`a, Yuki Mitsufuji 4/30/2026

Woosh: A Sound Effects Foundation Model

Woosh is Sony AI's open sound effects foundation model with audio encoder/decoder and text-to-audio generation capabilities.