7 papers · 1 filter
MERA: Model Evolution and Routing with Skill Adaptation for Agentic Systems at Scale
Yuhang Yao, Zeyu Wang, Wanyi Chen +8
LLM agents execute heterogeneous sequences of model calls within a single task: some invocations require careful reasoning, while others are structured steps such as formatting or…
Scaling Laws for Collapse in Asynchronous GRPO
Jingwei Song, Haofeng Xu, Jie Xiao +10
Asynchronous reinforcement learning improves the throughput of large language model post-training by decoupling rollout generation from policy optimization, but introduces a mismat…
STaR-KV: Spatio-Temporal Adaptive Re-weighting for KV Cache Compression in GUI Vision-Language Models
Yuhang Han, Wenzheng Yang, Yujie Chen +4
Vision-language-model-based graphical user interface (GUI) agents have shown broad automation capabilities, yet deployment is bottlenecked by a key-value (KV) cache that grows line…
TwinRouterBench: Fast Static and Live Dynamic Evaluation for Realistic Agentic LLM Routing
Pei Yang, Wanyi Chen, Tongyun Yang +15
LLM routing matters most in long-horizon applications such as coding agents, deep research systems, and computer-use agents, where a single user request triggers many model calls.…
FlexDraft: Flexible Speculative Decoding via Attention Tuning and Bonus-Guided Calibration
Yaojie Zhang, Jianuo Huang, Junlong Ke +5
Speculative decoding accelerates memory-bound LLM inference without quality degradation by using a fast drafter to propose multiple candidate tokens and the target model to verify…
Bridging Visual Representation and Reinforcement Learning from Verifiable Rewards in Large Vision-Language Models
Yuhang Han, Yuyang Wu, Zhengbo Jiao +6
Reinforcement Learning from Verifiable Rewards (RLVR) has substantially enhanced the reasoning capabilities of large language models in abstract reasoning tasks. However, its appli…
Innovator-VL: A Multimodal Large Language Model for Scientific Discovery
Zichen Wen, Boxue Yang, Shuang Chen +30
We present Innovator-VL, a scientific multimodal large language model designed to advance understanding and reasoning across diverse scientific domains while maintaining excellent…