22 papers
Chain-of-Experience for Continual LLM Improvement
Haoqin Tu, Yunhao Fang, Yizhong Wang +2
Humans continuously learn from experience, whereas conventional large language model (LLM) evaluations ignore the models' ability to improve through inference-time interaction. In…
ClawArena-Team: Benchmarking Subagent Orchestration and Dynamic Workflows in Language-Model Agents
Kaiwen Xiong, Haonian Ji, Shi Qiu +4
Production large language-model (LLM) agents are increasingly deployed not as lone problem-solvers but as managers: a main model creates specialized subagents, delegates work, and…
VisualClaw: A Real-Time, Personalized Agent for the Physical World
Haoqin Tu, Jianwen Chen, Zijun Wang +14
Vision language models are serving as general-purpose interfaces for complex multimodal tasks. However, deployment still faces three gaps: VLMs typically incur high latency and cos…
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
Minhua Lin, Juncheng Wu, Zijun Wang +14
LLM agents are increasingly deployed as systems built around editable external harnesses, including prompts, skills, memories and tools, that shape task execution without changing…
AutoResearchClaw: Self-Reinforcing Autonomous Research with Human-AI Collaboration
Jiaqi Liu, Shi Qiu, Mairui Li +33
Automating scientific discovery requires more than generating papers from ideas. Real research is iterative: hypotheses are challenged from multiple perspectives, experiments fail…
From Seeing to Thinking: Decoupling Perception and Reasoning Improves Post-Training of Vision-Language Models
Juncheng Wu, Hardy Chen, Haoqin Tu +6
Recent advances in vision-language models (VLMs) emphasize long chain-of-thought reasoning; yet, we find that their performance on visual tasks is primarily limited by a lack of vi…
ClinSeekAgent: Automating Multimodal Evidence Seeking for Agentic Clinical Reasoning
Juncheng Wu, Letian Zhang, Yuhan Wang +9
Large language models (LLMs) and agentic systems have shown promise for clinical decision support, but existing works largely assume that evidence has already been curated and hand…
ClawArena: Benchmarking AI Agents in Evolving Information Environments
Haonian Ji, Kaiwen Xiong, Siwei Han +9
AI agents deployed as persistent assistants must maintain correct beliefs as their information environment evolves. In practice, evidence is scattered across heterogeneous sources…
ClawMark: A Living-World Benchmark for Multi-Turn, Multi-Day, Multimodal Coworker Agents
Fanqing Meng, Lingxiao Du, Zijian Wu +46
Language-model agents are increasingly used as persistent coworkers that assist users across multiple working days. During such workflows, the surrounding environment may change in…
VLAA-GUI: Knowing When to Stop, Recover, and Search, A Modular Framework for GUI Automation
Qijun Han, Haoqin Tu, Zijun Wang +11
Autonomous GUI agents face two fundamental challenges: early stopping, where agents prematurely declare success without verifiable evidence, and repetitive loops, where agents cycl…
Chasing the Public Score: User Pressure and Evaluation Exploitation in Coding Agent Workflows
Hardy Chen, Nancy Lau, Haoqin Tu +8
Frontier coding agents are increasingly used in workflows where users supervise progress primarily through repeated improvement of a public score, namely the reported score on a pu…
On Safety Risks in Experience-Driven Self-Evolving Agents
Weixiang Zhao, Yichen Zhang, Yingshuo Wang +8
Experience-driven self-evolution has emerged as a promising paradigm for improving the autonomy of large language model agents, yet its reliance on self-curated experience introduc…
Target-Oriented Pretraining Data Selection via Neuron-Activated Graph
Zijun Wang, Haoqin Tu, Weidong Zhou +7
Everyday tasks come with a target, and pretraining models around this target is what turns them into experts. In this paper, we study target-oriented language model (LM) pretrainin…
Your Agent, Their Asset: A Real-World Safety Analysis of OpenClaw
Zijun Wang, Haoqin Tu, Letian Zhang +11
OpenClaw, the most widely deployed personal AI agent in early 2026, operates with full local system access and integrates with sensitive services such as Gmail, Stripe, and the fil…
Omni-SimpleMem: Autoresearch-Guided Discovery of Lifelong Multimodal Agent Memory
Jiaqi Liu, Zipeng Ling, Shi Qiu +9
AI agents increasingly operate over extended time horizons, yet their ability to retain, organize, and recall multimodal experiences remains a critical bottleneck. Building effecti…
Omni-MMSI: Toward Identity-attributed Social Interaction Understanding
Xinpeng Li, Bolin Lai, Hardy Chen +5
We introduce Omni-MMSI, a new task that requires comprehensive social interaction understanding from raw audio, vision, and speech input. The task involves perceiving identity-attr…
MetaClaw: Just Talk -- An Agent That Meta-Learns and Evolves in the Wild
Peng Xia, Jianwen Chen, Xinyu Yang +10
Large language model (LLM) agents are increasingly used for complex tasks, yet deployed agents often remain static, failing to adapt as user needs evolve. This creates a tension be…
Kestrel: Grounding Self-Refinement for LVLM Hallucination Mitigation
Jiawei Mao, Hardy Chen, Haoqin Tu +7
Large vision-language models (LVLMs) have become increasingly strong but remain prone to hallucinations in multimodal tasks, which significantly narrows their deployment. As traini…
SimpleMem: Efficient Lifelong Memory for LLM Agents
Jiaqi Liu, Yaofeng Su, Peng Xia +5
To support long-term interaction in complex environments, LLM agents require memory systems that manage historical experiences. Existing approaches either retain full interaction h…
VQ-VA World: Towards High-Quality Visual Question-Visual Answering
Chenhui Gou, Zilong Chen, Zeyu Wang +10
This paper studies Visual Question-Visual Answering (VQ-VA): generating an image, rather than text, in response to a visual question -- an ability that has recently emerged in prop…
When Visualizing is the First Step to Reasoning: MIRA, a Benchmark for Visual Chain-of-Thought
Yiyang Zhou, Haoqin Tu, Zijun Wang +11
We propose MIRA, a new benchmark designed to evaluate models in scenarios where generating intermediate visual images is essential for successful reasoning. Unlike traditional CoT…
Alignment Tipping Process: How Self-Evolution Pushes LLM Agents Off the Rails
Siwei Han, Kaiwen Xiong, Jiaqi Liu +9
As Large Language Model (LLM) agents increasingly gain self-evolutionary capabilities to adapt and refine their strategies through real-world interaction, their long-term reliabili…