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
Jiacheng Liu, Taylor Blanton, Yanai Elazar +28 authors
OLMoTrace traces language model outputs to their training data in real-time, providing insights into fact checking, hallucination, and creativity through verbatim matches and infini-gram technology.
Weiye Xu, Jiahao Wang, Weiyun Wang +10 authors
VisuLogic is a benchmark to evaluate genuine vision-centric reasoning in multimodal large language models, revealing significant performance gaps compared to human accuracy.
Cheng Yang, Haiyuan Wan, Yiran Peng +8 authors
VR-Bench evaluates video models' spatial reasoning capabilities through maze-solving tasks, demonstrating that these models excel in spatial perception and reasoning, outperforming VLMs and benefiting from diverse sampling during inference.
Jiahan Zhang, Muqing Jiang, Nanru Dai +14 authors
World-in-World evaluates generative world models in closed-loop environments, emphasizing task success over visual quality and revealing insights into controllability, data scaling, and compute allocation.
Long Phan, Alice Gatti, Ziwen Han +661 authors
HLE is a challenging multi-modal benchmark that highlights the limitations of current LLMs in closed-ended academic questions.
Zhiheng Liu, Weiming Ren, Haozhe Liu +22 authors
TUNA, a unified multimodal model, uses a cascaded VAE and representation encoder for end-to-end multimodal understanding and generation, outperforming decoupled models and achieving state-of-the-art results across various benchmarks.
Marianne Arriola, Aaron Gokaslan, Justin T Chiu +5 authors
Block diffusion language models improve generation efficiency and sequence length flexibility compared to autoregressive and discrete denoising diffusion models.
Shuai Wang, Zhi Tian, Weilin Huang +1 authors
A decoupled diffusion transformer improves performance and training speed in image generation by separating semantic extraction and high-frequency decoding.
Yang Sui, Yu-Neng Chuang, Guanchu Wang +9 authors
This survey examines progress and methods for improving efficient reasoning in Large Language Models (LLMs), focusing on model, output, and input prompt optimizations, as well as the use of efficient data and small models.
Xingtai Lv, Yuxin Zuo, Youbang Sun +9 authors
A unified policy gradient estimator and Hybrid Post-Training algorithm effectively combine online and offline data for post-training language models, improving performance across various benchmarks.
Zhenglin Cheng, Peng Sun, Jianguo Li +1 authors
TwinFlow is a 1-step generative model framework that enhances inference efficiency without requiring fixed pretrained teacher models or standard adversarial networks, achieving high performance on text-to-image tasks and scaling efficiently.
Jintao Zhang, Jia Wei, Pengle Zhang +6 authors
Efficiency enhancements for attention mechanisms, including leveraging FP4 Tensor Cores and developing an 8-bit attention method, improve inference and training performance.
Shengqiong Wu, Weicai Ye, Jiahao Wang +8 authors
Any2Caption framework improves video generation quality and controllability by decoupling condition interpretation and leveraging multimodal large language models.
Boyong Wu, Chao Yan, Chen Hu +106 authors
Step-Audio~2, an end-to-end multi-modal large language model, integrates latent audio encoding and reinforcement learning to achieve state-of-the-art performance in ASR, audio understanding, and speech conversation, incorporating discrete audio token generation and retrieval-augmented generation.
Xiaopeng Lin, Shijie Lian, Bin Yu +8 authors
Proposed Egocentric2Embodiment pipeline translates human egocentric videos into structured training data for robots, enhancing their egocentric understanding and task performance.
Itay Nakash, Nitay Calderon, Eyal Ben David +2 authors
AdaptiVocab, a novel vocabulary adaptation method, enhances the efficiency of Large Language Models in domain-specific applications without sacrificing performance by reducing token usage.
Jianxiong Gao, Zhaoxi Chen, Xian Liu +7 authors
LongVie 2, an end-to-end autoregressive framework, enhances controllability, visual quality, and temporal consistency in video world models through three progressive training stages.
Wei Shen, Jiangbo Pei, Yi Peng +7 authors
Skywork-R1V3, an open-source vision-language model, enhances visual reasoning through a post-training reinforcement learning framework, achieving state-of-the-art performance on multimodal reasoning tasks.
Juyuan Wang, Rongchen Zhao, Wei Wei +5 authors
ComoRAG, an iterative retrieval-based approach, enhances long-context narrative comprehension by dynamically updating memory and generating probing queries, outperforming traditional RAG methods.
Yana Wei, Liang Zhao, Jianjian Sun +15 authors
A two-stage paradigm using Qwen2.5-VL-7B and multimodal reinforcement learning achieves state-of-the-art performance in visual reasoning benchmarks.
Yuxiang Wei, Olivier Duchenne, Jade Copet +6 authors
SWE-RL leverages reinforcement learning to enhance large language models' software engineering reasoning, achieving state-of-the-art performance and generalized reasoning skills by learning from extensive open-source software evolution data.
Xiaochuan Li, Guoguang Du, Runze Zhang +11 authors
Using video data to provide commonsense priors enhances 3D asset generation, enabling spatial consistency and semantic plausibility in 3D content creation.
Yuhao Zhou, Yiheng Wang, Xuming He +24 authors
Scientists' First Exam (SFE) benchmark assesses scientific cognitive capacities of Multimodal Large Language Models through perception, understanding, and comparative reasoning.
Nan Chen, Mengqi Huang, Yihao Meng +1 authors
LongAnimation framework achieves long-term color consistency in animation colorization by integrating a dynamic global-local memory and color consistency reward.
Ke Xing, Longfei Li, Yuyang Yin +8 authors
StereoWorld generates high-quality stereo video from monocular input using a pretrained video generator with geometry-aware regularization and spatio-temporal tiling.
Soyeong Jeong, Kangsan Kim, Jinheon Baek +1 authors
VideoRAG dynamically retrieves relevant videos based on queries and utilizes both visual and textual information for output generation, outperforming existing methods in response accuracy.
Xiangyu Zhao, Shengyuan Ding, Zicheng Zhang +10 authors
OmniAlign-V improves human preference alignment in MLLMs through supervised fine-tuning and preference optimization without compromising performance on VQA benchmarks.
Ziyi Xia, Kun Luo, Hongjin Qian +1 authors
InfoSeek is a scalable framework for generating complex Deep Research tasks by synthesizing hierarchical constraint satisfaction problems, enabling models to outperform larger baselines on challenging benchmarks.
Zhenran Xu, Longyue Wang, Jifang Wang +7 authors
FilmAgent, a multi-agent collaborative framework based on LLMs, automates film production processes including scriptwriting, cinematography, and actor positioning, outperforming other models in human evaluations.
Yuri Kuratov, Mikhail Arkhipov, Aydar Bulatov +1 authors
Using per-sample optimization, compression ratios of up to x1500 are achieved for sequence-to-vector compression in language models, highlighting a large gap between theoretical and practical limits.
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