8 papers
A Distribution-Matching Perspective on LLM Training Algorithms
Yanxing Huang, Bingzhang Yan, Zihan Tang +3
In this work, we propose a unified framework for training algorithms across different stages of modern large language model development, based on the perspective of distribution ma…
Orca: The World is in Your Mind
Yihao Wang, Yuheng Ji, Mingyu Cao +54
We introduce Orca, an initial instantiation of a general world foundation model. Orca learns a unified world latent space from multimodal world signals and exposes it through multi…
From Reasoning Chains to Verifiable Subproblems: Curriculum Reinforcement Learning Enables Credit Assignment for LLM Reasoning
Xitai Jiang, Zihan Tang, Wenze Lin +3
Reinforcement learning from verifiable rewards (RLVR) has shown strong promise for LLM reasoning, but outcome-based RLVR remains inefficient on hard problems because correct final-…
RTPrune: Reading-Twice Inspired Token Pruning for Efficient DeepSeek-OCR Inference
Ben Wan, Yan Feng, Zihan Tang +4
DeepSeek-OCR leverages visual-text compression to reduce long-text processing costs and accelerate inference, yet visual tokens remain prone to redundant textual and structural inf…
FastOCR: Dynamic Visual Fixation via KV Cache Pruning for Efficient Document Parsing
Zihan Tang, Leqi Shen, Hui Chen +7
Vision-Language Models (VLMs) have shown strong promise on Optical Character Recognition (OCR), yet the sheer number of visual tokens required to encode dense documents incurs proh…
OOCO: Latency-disaggregated Architecture for Online-Offline Co-locate LLM Serving
Siyu Wu, Zihan Tang, Yuting Zeng +5
Large Language Models (LLMs) are increasingly deployed in both latency-sensitive online services and cost-sensitive offline workloads. Co-locating these workloads on shared serving…
Pessimistic Verification for Open Ended Math Questions
Yanxing Huang, Zihan Tang, Zejin Lin +2
Automatic verification is a critical component in building math-solving agents and reinforcement learning, yet it often falls short in generalizability, performance, and cost-effic…
xLLM Technical Report
Tongxuan Liu, Tao Peng, Peijun Yang +50
We introduce xLLM, an intelligent and efficient Large Language Model (LLM) inference framework designed for high-performance, large-scale enterprise-grade serving, with deep optimi…