R. Xu
YOU?
Author Swipe
View article: Unifying Language Agent Algorithms with Graph-based Orchestration Engine for Reproducible Agent Research
Unifying Language Agent Algorithms with Graph-based Orchestration Engine for Reproducible Agent Research Open
Language agents powered by large language models (LLMs) have demonstrated remarkable capabilities in understanding, reasoning, and executing complex tasks. However, developing robust agents presents significant challenges: substantial engi…
View article: The Self-Improvement Paradox: Can Language Models Bootstrap Reasoning Capabilities without External Scaffolding?
The Self-Improvement Paradox: Can Language Models Bootstrap Reasoning Capabilities without External Scaffolding? Open
Self-improving large language models (LLMs) -- i.e., to improve the performance of an LLM by fine-tuning it with synthetic data generated by itself -- is a promising way to advance the capabilities of LLMs while avoiding extensive supervis…
View article: ZoomEye: Enhancing Multimodal LLMs with Human-Like Zooming Capabilities through Tree-Based Image Exploration
ZoomEye: Enhancing Multimodal LLMs with Human-Like Zooming Capabilities through Tree-Based Image Exploration Open
Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in vision-language understanding. Recently, with the integration of test-time scaling techniques, these models have also shown strong potential in visual re…
View article: OmChat: A Recipe to Train Multimodal Language Models with Strong Long Context and Video Understanding
OmChat: A Recipe to Train Multimodal Language Models with Strong Long Context and Video Understanding Open
We introduce OmChat, a model designed to excel in handling long contexts and video understanding tasks. OmChat's new architecture standardizes how different visual inputs are processed, making it more efficient and adaptable. It uses a dyn…
View article: Preserving Knowledge in Large Language Model with Model-Agnostic Self-Decompression
Preserving Knowledge in Large Language Model with Model-Agnostic Self-Decompression Open
Humans can retain old knowledge while learning new information, but Large Language Models (LLMs) often suffer from catastrophic forgetting when post-pretrained or supervised fine-tuned (SFT) on domain-specific data. Moreover, for Multimoda…
View article: Numerical Simulation of Inlet Void Fraction Affecting Oil-gas Two-phase Flow Characteristics in 90° Elbows
Numerical Simulation of Inlet Void Fraction Affecting Oil-gas Two-phase Flow Characteristics in 90° Elbows Open
Air can have an adverse effect on the performance of an aero-engine lubrication system. A numerical analysis was conducted to explore the influence of inlet void fraction and pipe layout on the characteristics of oil-gas two-phase flow in …
View article: First Search for Light Fermionic Dark Matter Absorption on Electrons Using Germanium Detector in CDEX-10 Experiment
First Search for Light Fermionic Dark Matter Absorption on Electrons Using Germanium Detector in CDEX-10 Experiment Open
We present the first results of the search for sub-MeV fermionic dark matter absorbed by electron targets of Germanium using the 205.4~kg$\cdot$day data collected by the CDEX-10 experiment, with the analysis threshold of 160~eVee. No signi…