14 papers
ATTUNER: Recomputation-Free Agentic Context KV Reuse
Xinghao Chen, Junnan Dong, Cai Ke +8
Large language model (LLM) agents repeatedly load reusable content, such as skills, documents, and memory entries, into the current context. Re-encoding this content for every requ…
Latent Multimodal Memory Reasoning
Cai Ke, Xinghao Chen, Yuanchen Bei +7
Long-term personalized agents are required to reason over distributed history across modalities and evolving dialogue states. While existing flat cross-modal methods compress visua…
Draft in Parallel, Condition Through Depth: Adjacent Causal Injection for Speculative Decoding
Haohui Zhang, Keyu Chen, Haocheng Sun +4
Parallel speculative drafting generates multiple candidates in one backbone pass, but independent token selection can produce inconsistent continuations that shorten the accepted p…
From Atomic to Agentic: Towards Interpretable Evaluation of LLMs' Agentic Mathematical Capabilities
Jiayi Kuang, Yinghui Li, Yunze Song +11
Large Language Models (LLMs) are evolving from performing end-to-end mathematical reasoning to integrating agentic intelligence. However, most existing math benchmarks evaluate onl…
Training-Free Hashing-Based Attention via Binary Principal Components
Daohai Yu, Zhanpeng Zeng, Keyu Chen +6
Long-context large language models (LLMs) are increasingly deployed in real-world applications, yet self-attention remains a major efficiency bottleneck -- especially during decodi…
ManimAgent: Self-Evolving Multimodal Agents for Visual Education
Wenjia Jiang, Zongyuan Cai, Yuanhang Shao +7
Multi-round reflection lets agents built on large language models recover from failures within a single task, but each task remains an isolated episode: lessons learned across many…
Skill Is Not Document: Query-Conditioned Compatibility for LLM Agent Skill Routing
Zifei Wang, Wei Wen, Qiang Ji +3
Large language model agents increasingly rely on reusable skills, making skill retrieval a critical front-end component of agent systems. Skill retrieval, however, is not ordinary…
Toward Native Multimodal Modeling: A Roadmap
Siyu An, Junru Lu, Junnan Dong +18
Multimodal modeling represents a vital step from modality-agnostic reasoning toward world modeling. While early approaches predominantly rely on late-fusion that assembles encoders…
Benchmarking and Evolving Reason-Reflect-Rectify for Reflective Visual Generation
Junjie Wang, Xinghua Lou, Jason Li +8
Text-to-Image (T2I) models and Unified Multimodal Models (UMMs) have achieved remarkable progress in visual generation. However, their reliance on a single-pass generation paradigm…
When should I search more: Adaptive Complex Query Optimization with Reinforcement Learning
Wei Wen, Sihang Deng, Tianjun Wei +3
Query optimization is a crucial component for the efficacy of Retrieval-Augmented Generation (RAG) systems. While reinforcement learning (RL)-based agentic and reasoning methods ha…
Youtu-LLM: Unlocking the Native Agentic Potential for Lightweight Large Language Models
Junru Lu, Jiarui Qin, Lingfeng Qiao +35
We introduce Youtu-LLM, a lightweight yet powerful language model that harmonizes high computational efficiency with native agentic intelligence. Unlike typical small models that r…
Adaptive Dual Reasoner: Large Reasoning Models Can Think Efficiently by Hybrid Reasoning
Yujian Zhang, Keyu Chen, Zhifeng Shen +2
Although Long Reasoning Models (LRMs) have achieved superior performance on various reasoning scenarios, they often suffer from increased computational costs and inference latency…
HiChunk: Evaluating and Enhancing Retrieval-Augmented Generation with Hierarchical Chunking
Wensheng Lu, Keyu Chen, Ruizhi Qiao +1
Retrieval-Augmented Generation (RAG) enhances the response capabilities of language models by integrating external knowledge sources. However, document chunking as an important par…
ASPD: Unlocking Adaptive Serial-Parallel Decoding by Exploring Intrinsic Parallelism in LLMs
Keyu Chen, Zhifeng Shen, Daohai Yu +5
The increasing scale and complexity of large language models (LLMs) pose significant inference latency challenges, primarily due to their autoregressive decoding paradigm character…