10 papers
PAC: Progress-Augmented Advantage Curriculum for Multi-Task Reinforcement Learning of LLMs
Yuanqiang Yu, Yanzhao Zheng, Zhentao Zhang +8
Reinforcement learning (RL) is used to improve the reasoning abilities of LLMs, while training data span heterogeneous tasks. However, most RL post-training pipelines rely on fixed…
Which Tokens Matter? Adaptive Token Selection for RLVR with the Relative Surprisal Index
Outongyi Lv, Yanzhao Zheng, Yuanwei Zhang +5
Reinforcement learning (RL) has become a powerful tool for propelling Large Language Models (LLMs) beyond imitation-based training towards more robust reasoning capabilities. Among…
Rethinking Memory as Continuously Evolving Connectivity
Jizhan Fang, Buqiang Xu, Zhixian Wang +12
Existing memory-augmented LLM agents often treat memory as a static repository with pre-defined representations and fixed retrieval pipelines, which is brittle in dynamic agentic e…
MemTrace: Tracing and Attributing Errors in Large Language Model Memory Systems
Xinle Deng, Ruobin Zhong, Hujin Peng +15
Memory is essential for enabling large language models to support long-horizon reasoning, yet existing memory systems remain unreliable and difficult to debug. Tracing memory's dyn…
Rubrics to Tokens: Bridging Response-level Rubrics and Token-level Rewards in Instruction Following Tasks
Tianze Xu, Yanzhao Zheng, Pengrui Lu +11
Rubric-based Reinforcement Learning (RL) has emerged as a promising approach for aligning Large Language Models (LLMs) with complex, open-domain instruction following tasks. Howeve…
ContextBudget: Budget-Aware Context Management for Long-Horizon Search Agents
Yong Wu, YanZhao Zheng, TianZe Xu +9
LLM-based agents show strong potential for long-horizon reasoning, yet their context size is limited by deployment factors (e.g., memory, latency, and cost), yielding a constrained…
SkillRouter: Skill Routing for LLM Agents at Scale
YanZhao Zheng, ZhenTao Zhang, Chao Ma +8
Reusable skills let LLM agents package task-specific procedures, tool affordances, and execution guidance into modular building blocks. As skill ecosystems grow to tens of thousand…
SERL: Self-Examining Reinforcement Learning on Open-Domain
Weixuan Ou, Yanzhao Zheng, Shuoshuo Sun +7
Reinforcement Learning (RL) has been shown to improve the capabilities of large language models (LLMs). However, applying RL to open-domain tasks faces two key challenges: (1) the…
Reason from Future: Reverse Thought Chain Enhances LLM Reasoning
Yinlong Xu, Yanzhao Zheng, Shuoshuo Sun +7
It has been demonstrated that carefully designed reasoning paradigms, like Chain-of-Thought (CoT) and Tree-of-Thought (ToT), can enhance the reasoning capabilities of small languag…
TextMaster: A Unified Framework for Realistic Text Editing via Glyph-Style Dual-Control
Zhenyu Yan, Jian Wang, Aoqiang Wang +3
In image editing tasks, high-quality text editing capabilities can significantly reduce both human and material resource costs. Existing methods, however, face significant limitati…