15 papers
Scaling GUI Agents with Visual State Transitions
Xiangyan Liu, Kaixin Li, Haonan Wang +6
We introduce State Transition Pretraining (STP) as a new scaling axis for GUI agents. During the STP stage, we continually pretrain a unified multimodal model on visual state trans…
Multimodal Unlearning Across Vision, Language, Video, and Audio: Survey of Methods, Datasets, and Benchmarks
Nobin Sarwar, Shubhashis Roy Dipta, Zheyuan Liu +1
With the growing adoption of VLMs, DMs, LLMs, and AFMs, these multimodal foundation models can inadvertently encode sensitive, copyrighted, biased, or unsafe cross-modal associatio…
SAVAA: Mitigating Hallucinations in LVLMs via Step-wise Adaptive Visual Attention Amplification
Jiacheng Zhang, Feng Liu, Chao Du +1
A line of recent training-free methods for mitigating hallucinations in large vision-language models (LVLMs) operates by amplifying attention to visual tokens during autoregressive…
Demystifying the Slash Pattern in Attention: The Role of RoPE
Yuan Cheng, Fengzhuo Zhang, Yunlong Hou +5
Large Language Models (LLMs) often exhibit slash attention patterns, where attention scores concentrate along the -th sub-diagonal for some offset . These patterns play a key…
Think in Parallel, Answer as One: Logit Averaging for Open-Ended Reasoning
Haonan Wang, Chao Du, Kenji Kawaguchi +1
Majority voting has proven effective for close-ended question answering by aggregating parallel reasoning traces. However, it is not directly applicable to open-ended reasoning, su…
Diffusion Language Models are Super Data Learners
Jinjie Ni, Qian Liu, Longxu Dou +5
Under strictly controlled pre-training settings, we observe a Crossover: when unique data is limited, diffusion language models (DLMs) consistently surpass autoregressive (AR) mode…
Training Optimal Large Diffusion Language Models
Jinjie Ni, Qian Liu, Chao Du +5
We introduce Quokka, the first systematic scaling law for diffusion language models (DLMs), encompassing both compute-constrained and data-constrained regimes, and studying the key…
VerlTool: Towards Holistic Agentic Reinforcement Learning with Tool Use
Dongfu Jiang, Yi Lu, Zhuofeng Li +9
Reinforcement Learning with Verifiable Rewards (RLVR) has demonstrated success in enhancing LLM reasoning capabilities, but remains limited to single-turn interactions without tool…
Muon Outperforms Adam in Tail-End Associative Memory Learning
Shuche Wang, Fengzhuo Zhang, Jiaxiang Li +6
The Muon optimizer is consistently faster than Adam in training Large Language Models (LLMs), yet the mechanism underlying its success remains unclear. This paper demystifies this…
Fostering Video Reasoning via Next-Event Prediction
Haonan Wang, Hongfu Liu, Xiangyan Liu +4
Next-token prediction serves as the foundational learning task enabling reasoning in LLMs. But what should the learning task be when aiming to equip MLLMs with temporal reasoning c…
Adversarial Attacks against Closed-Source MLLMs via Feature Optimal Alignment
Xiaojun Jia, Sensen Gao, Simeng Qin +7
Multimodal large language models (MLLMs) remain vulnerable to transferable adversarial examples. While existing methods typically achieve targeted attacks by aligning global featur…
Reinforcing General Reasoning without Verifiers
Xiangxin Zhou, Zichen Liu, Anya Sims +6
The recent paradigm shift towards training large language models (LLMs) using DeepSeek-R1-Zero-style reinforcement learning (RL) on verifiable rewards has led to impressive advance…
Lifelong Safety Alignment for Language Models
Haoyu Wang, Zeyu Qin, Yifei Zhao +4
LLMs have made impressive progress, but their growing capabilities also expose them to highly flexible jailbreaking attacks designed to bypass safety alignment. While many existing…
FlowReasoner: Reinforcing Query-Level Meta-Agents
Hongcheng Gao, Yue Liu, Yufei He +6
This paper proposes a query-level meta-agent named FlowReasoner to automate the design of query-level multi-agent systems, i.e., one system per user query. Our core idea is to ince…
NoisyRollout: Reinforcing Visual Reasoning with Data Augmentation
Xiangyan Liu, Jinjie Ni, Zijian Wu +5
Recent advances in reinforcement learning (RL) have strengthened the reasoning capabilities of vision-language models (VLMs). However, enhancing policy exploration to better scale…