Explore · 每周精选
发现最受关注的研究论文,追踪研究趋势,订阅感兴趣的期刊与关键词。
606 篇论文 · 按点赞排序
Kristian Kuznetsov, Laida Kushnareva, Polina Druzhinina +5 authors
Sparse Autoencoders enhance interpretability in Artificial Text Detection by extracting distinctive features from LLM outputs, providing insights into differences from human-written texts.
Rui-Jie Zhu, Zixuan Wang, Kai Hua +30 authors
LoopLM, a family of pre-trained Looped Language Models, enhances reasoning by integrating iterative computation and entropy regularization during pre-training, achieving superior performance with better knowledge manipulation.
Hao Liang, Xiaochen Ma, Zhou Liu +32 authors
DataFlow is an LLM-driven data preparation framework that enhances data quality and reproducibility for various tasks, improving LLM performance with automatically generated pipelines.
Gaojie Lin, Jianwen Jiang, Jiaqi Yang +2 authors
OmniHuman, a Diffusion Transformer-based framework, achieves realistic human video generation by integrating motion-related conditions into its training process, supporting various inputs and driving modalities.
Weiyun Wang, Zhangwei Gao, Lixin Gu +58 authors
InternVL 3.5 introduces Cascade RL, ViR, and DvD to enhance reasoning, efficiency, and performance in multimodal models.
Khalil Hennara, Muhammad Hreden, Mohamed Motaism Hamed +3 authors
Mutarjim is a compact Arabic-English translation model that outperforms larger models on established benchmarks and achieves state-of-the-art performance on a new comprehensive Tarjama-25 benchmark.
Shuai Bai, Keqin Chen, Xuejing Liu +24 authors
Qwen2.5-VL, the latest vision-language model, advances visual recognition, document parsing, and video comprehension through dynamic resolution processing, Window Attention, and a native Vision Transformer.
Weihao Tan, Xiangyang Li, Yunhao Fang +11 authors
Lumine, a vision-language model-based agent, completes complex missions in real-time across different 3D open-world environments with human-like efficiency and zero-shot cross-game generalization.
Yang Wang, Chenghao Xiao, Chia-Yi Hsiao +4 authors
LLMs struggle with understanding the nuanced, context-dependent meanings of Drivelological text, which appears nonsensical but contains deeper semantic layers.
GLM-4. 5 Team, Aohan Zeng, Xin Lv +168 authors
GLM-4.5, a Mixture-of-Experts large language model with 355B parameters, achieves strong performance across agentic, reasoning, and coding tasks using multi-stage training and reinforcement learning.
Andrés Marafioti, Orr Zohar, Miquel Farré +14 authors
SmolVLM, a series of compact multimodal models, achieves high performance with minimal GPU memory usage, making efficient deployment on mobile and edge devices possible.
Adam Tauman Kalai, Ofir Nachum, Santosh S. Vempala +1 authors
Language models produce incorrect statements due to training and evaluation procedures that reward guessing over acknowledging uncertainty, leading to a need for socio-technical changes in benchmark scoring.
MiroMind Team, Song Bai, Lidong Bing +51 authors
We present MiroThinker v1.0, an open-source research agent designed to advance tool-augmented reasoning and information-seeking capabilities. Unlike previous agents that only scale up model size or context length, MiroThinker explores interaction scaling at the model level, systematically training the model to handle deeper and more frequent agent-environment interactions as a third dimension of performance improvement. Unlike LLM test-time scaling, which operates in isolation and risks degradation with longer reasoning chains, interactive scaling leverages environment feedback and external information acquisition to correct errors and refine trajectories. Through reinforcement learning, the model achieves efficient interaction scaling: with a 256K context window, it can perform up to 600 tool calls per task, enabling sustained multi-turn reasoning and complex real-world research workflows. Across four representative benchmarks-GAIA, HLE, BrowseComp, and BrowseComp-ZH-the 72B variant achieves up to 81.9%, 37.7%, 47.1%, and 55.6% accuracy respectively, surpassing previous open-source agents and approaching commercial counterparts such as GPT-5-high. Our analysis reveals that MiroThinker benefits from interactive scaling consistently: research performance improves predictably as the model engages in deeper and more frequent agent-environment interactions, demonstrating that interaction depth exhibits scaling behaviors analogous to model size and context length. These findings establish interaction scaling as a third critical dimension for building next-generation open research agents, complementing model capacity and context windows.
Deepak Nathani, Lovish Madaan, Nicholas Roberts +14 authors
A new framework and benchmark, MLGym and MLGym-Bench, is introduced for evaluating and developing LLM agents across diverse AI research tasks using reinforcement learning.
Mo Yu, Lemao Liu, Junjie Wu +5 authors
A study investigates whether large language models understand physical concepts through a grid-based task, showing that they significantly underperform compared to humans and highlighting the stochastic parrot phenomenon.
Kaiyan Zhang, Yuxin Zuo, Bingxiang He +36 authors
Reinforcement Learning enhances Large Language Models for complex reasoning tasks, facing challenges in scalability and infrastructure as the field advances.
Andrew Zhao, Yiran Wu, Yang Yue +8 authors
Absolute Zero Reasoner (AZR) achieves state-of-the-art performance on coding and mathematical reasoning tasks through self-generated, verified learning without external data.
Shenzhi Wang, Le Yu, Chang Gao +15 authors
Token entropy patterns are crucial in RLVR, with high-entropy tokens significantly impacting reasoning performance and model optimization.
Yongliang Wu, Yizhou Zhou, Zhou Ziheng +7 authors
Dynamic Fine-Tuning (DFT) improves the generalization of Large Language Models (LLMs) by dynamically rescaling gradients, outperforming standard Supervised Fine-Tuning (SFT) and showing competitive results in offline reinforcement learning.
Shuai Yang, Wei Huang, Ruihang Chu +9 authors
LongLive is a frame-level autoregressive framework for real-time and interactive long video generation, addressing efficiency and quality challenges through causal attention, KV-recache, streaming long tuning, and short window attention.
Zuhao Yang, Sudong Wang, Kaichen Zhang +7 authors
LongVT, an end-to-end framework, enhances long video reasoning by interleaving global and local analysis using multimodal tools, outperforming existing methods on challenging benchmarks.
Yunxin Li, Zhenyu Liu, Zitao Li +19 authors
This survey outlines the development of multimodal reasoning models, progressing from task-specific modules to unified frameworks, and discusses emerging capabilities and challenges in integrating reasoning across diverse data types.
Yiying Yang, Wei Cheng, Sijin Chen +6 authors
OmniSVG, using pre-trained Vision-Language Models and discrete tokenization, generates high-quality SVGs efficiently with a large, multimodal dataset.
Wei Huang, Yi Ge, Shuai Yang +11 authors
QeRL, a quantization-enhanced reinforcement learning framework, accelerates RL training for large language models by combining NVFP4 quantization with Low-Rank Adaptation and an Adaptive Quantization Noise mechanism, achieving significant speedups and improved performance.
Yujia Zhang, Xiaoyang Wu, Yixing Lao +4 authors
Concerto, a minimalist model combining 3D self-distillation and 2D-3D joint embedding, achieves superior spatial feature learning and outperforms existing models in scene understanding and open-world perception.
Zijian Wu, Xiangyan Liu, Xinyuan Zhang +12 authors
MCPMark is a comprehensive benchmark for evaluating MCP use in real-world workflows, featuring diverse tasks that require richer interactions with the environment, and reveals that current LLMs perform poorly on these tasks.
Yubo Huang, Hailong Guo, Fangtai Wu +8 authors
Live Avatar uses a 14-billion-parameter diffusion model with Timestep-forcing Pipeline Parallelism and Rolling Sink Frame Mechanism to achieve real-time, high-fidelity avatar generation.
Anton Razzhigaev, Matvey Mikhalchuk, Temurbek Rahmatullaev +4 authors
Analysis of Large Language Models reveals that even minor tokens are crucial for context, and introduces LLM-Microscope for assessing token-level nonlinearity and contextual memory.
Junbo Niu, Zheng Liu, Zhuangcheng Gu +58 authors
MinerU2.5, a 1.2B-parameter document parsing vision-language model, achieves state-of-the-art recognition accuracy with computational efficiency through a coarse-to-fine parsing strategy.
Zhiliang Peng, Jianwei Yu, Wenhui Wang +10 authors
VibeVoice synthesizes long-form multi-speaker speech using next-token diffusion and a highly efficient continuous speech tokenizer, achieving superior performance and fidelity.
北京市昌平区探索星信息技术及软件开发工作室
京ICP备2026059466号