Ax Vishnu Narayanan Anilkumar, Abhijith Sreesylesh Babu, Trieu Hai Vo, Mohankrishna Kolla, Alexander Cuneo 3/18/2026

Relationship-Aware Safety Unlearning for Multimodal LLMs

Framework for unlearning relational safety failures in multimodal LLMs where combinations of benign concepts become unsafe when linked by specific relations.

Ax Yulin Peng, Xinxin Zhu, Chenxing Wei, Nianbo Zeng, Leilei Wang, Ying Tiffany He, F. Richard Yu 3/18/2026

SAGE: Multi-Agent Self-Evolution for LLM Reasoning

SAGE framework: multi-agent reinforcement learning system for improving LLM reasoning without large human-labeled datasets, using self-play and closed-loop feedback.

Ax Sebastian Stober, Tim W. Dornis 3/18/2026

Generative AI Training and Copyright Law

Interdisciplinary study on copyright law implications of training generative AI via web scraping, covering fair use and TDM exceptions.

Ax Donato Crisostomi, Alessandro Zirilli, Antonio Andrea Gargiulo, Maria Sofia Bucarelli, Simone Scardapane, Fabrizio Silvestri, Iacopo Masi, Emanuele Rodol\`a 3/18/2026

MASS: MoErging through Adaptive Subspace Selection

MASS method merges multiple fine-tuned models via adaptive subspace selection, improving accuracy over existing merging approaches without retraining.

Ax Eleonora Cappuccio (Department of Computer Science, University of Pisa), Andrea Esposito (Department of Computer Science, University of Bari Aldo Moro), Francesco Greco (Department of Computer Science, University of Bari Aldo Moro), Giuseppe Desolda (Department of Computer Science, University of Bari Aldo Moro), Rosa Lanzilotti (Department of Computer Science, University of Bari Aldo Moro), Salvatore Rinzivillo (ISTI CNR) 3/18/2026

Explanation User Interfaces: A Systematic Literature Review

Systematic literature review of explanation user interfaces for interpretable AI systems.

Ax Zhe Ye, Zhengxu Yan, Jingxuan He, Timothe Kasriel, Kaiyu Yang, Dawn Song 3/18/2026

VERINA: Benchmarking Verifiable Code Generation

VERINA benchmark for evaluating LLM code generation with joint code, specification, and proof generation.

Ax Weihua Du, Hailei Gong, Zhan Ling, Kang Liu, Lingfeng Shen, Xuesong Yao, Yufei Xu, Dingyuan Shi, Yiming Yang, Jiecao Chen 3/18/2026

Generalizable End-to-End Tool-Use RL with Synthetic CodeGym

CodeGym: Reinforcement learning framework for training LLM agents to use tools generalizing across new tasks and workflows.

Ax L\'eo Boisvert, Abhay Puri, Chandra Kiran Reddy Evuru, Nazanin Sepahvand, Nicolas Chapados, Quentin Cappart, Alexandre Lacoste, Krishnamurthy Dj Dvijotham, Alexandre Drouin 3/18/2026

Malice in Agentland: Down the Rabbit Hole of Backdoors in the AI Supply Chain

Security research identifying backdoor vulnerabilities in AI agent supply chains through data poisoning across multiple pipeline stages.

Ax Prajit Bhaskaran, Tom Viering 3/18/2026

Transformers can do Bayesian Clustering

Transformer-based model extending Prior-Data Fitted Networks for Bayesian clustering with uncertainty quantification on synthetic data.