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。
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.
50 篇论文 · 按点赞排序
Xuyang Liu, Zichen Wen, Shaobo Wang +13 authors
The focus in AI research is shifting from model-centric to data-centric compression, with token compression identified as key to improving efficiency in handling long-context scenarios.
Ganqu Cui, Yuchen Zhang, Jiacheng Chen +14 authors
Entropy dynamics in reinforcement learning with large language models are investigated to prevent policy entropy collapse and improve exploration.
Alan Arazi, Eilam Shapira, Roi Reichart
TabSTAR, a tabular foundation model with semantically target-aware representations, achieves state-of-the-art performance in classification tasks with text features through transfer learning without dataset-specific parameters.
Wei Pang, Kevin Qinghong Lin, Xiangru Jian +2 authors
A benchmark and metric suite for poster generation evaluates visual quality, coherence, and content accuracy, leading to a multi-agent pipeline that outperforms existing models with reduced computational cost.
Ibragim Badertdinov, Alexander Golubev, Maksim Nekrashevich +6 authors
A novel pipeline extracts real-world, interactive software engineering tasks from GitHub to create SWE-rebench, improving the evaluation of reinforcement learning models in SWE.
Zheyuan Yang, Lyuhao Chen, Arman Cohan +1 authors
Two post-training strategies, distillation and RLVR, enable inference-time scaling in table reasoning tasks, resulting in a model (Table-R1-Zero) that matches GPT-4.1's performance using fewer parameters and shows strong generalization.
Fanqi Wan, Weizhou Shen, Shengyi Liao +7 authors
A framework called QwenLong-L1 enhances large reasoning models for long-context reasoning through reinforcement learning, achieving leading performance on document question-answering benchmarks.
Jiakang Yuan, Tianshuo Peng, Yilei Jiang +8 authors
MME-Reasoning evaluates the logical reasoning capabilities of multimodal large language models, revealing significant limitations and performance imbalances across inductive, deductive, and abductive reasoning types.
Valerii Startsev, Alexander Ustyuzhanin, Alexey Kirillov +2 authors
A new method using a pre-trained generative model helps construct a high-impact SFT dataset, Alchemist, which improves the generative quality of text-to-image models while maintaining diversity.
Minki Kang, Jongwon Jeong, Seanie Lee +2 authors
Agent Distillation transfers reasoning and task-solving capabilities from large language models to smaller models using enhanced prompts and self-consistent actions, matching performance of larger models on various reasoning tasks.
Roberto L. Castro, Andrei Panferov, Soroush Tabesh +5 authors
Quartet, a hardware-supported FP4 training approach for large language models, demonstrates state-of-the-art accuracy while significantly reducing computational costs compared to standard or FP8 precision.
Tianyu Fu, Yi Ge, Yichen You +6 authors
Roads to Rome (R2R) selectively utilizes large language models for critical reasoning tasks to enhance efficiency and performance in lightweight models.
Diankun Wu, Fangfu Liu, Yi-Hsin Hung +1 authors
Spatial-MLLM improves spatial reasoning in multimodal large language models using a dual-encoder architecture with pretrained 2D and 3D structure encoders, achieving state-of-the-art performance on visual spatial tasks.
Junteng Liu, Yuanxiang Fan, Zhuo Jiang +12 authors
SynLogic, a data synthesis framework, enhances the logical reasoning capabilities of Large Language Models through RL, achieving state-of-the-art performance and improving generalization across various domains.
Yaorui Shi, Sihang Li, Chang Wu +5 authors
AutoRefine, a reinforcement learning framework for large language models, enhances retrieval-augmented reasoning by iteratively refining knowledge and optimizing searches, leading to improved performance in complex question-answering tasks.
Guilong Lu, Xuntao Guo, Rongjunchen Zhang +2 authors
BizFinBench is a benchmark for evaluating large language models in financial applications, revealing distinct performance patterns across various tasks.
Doohyuk Jang, Yoonjeon Kim, Chanjae Park +2 authors
A diagnostic set examines and categorizes reasoning rigidity in large language models, identifying patterns where models ignore instructions and default to familiar reasoning.
Yan Ma, Linge Du, Xuyang Shen +7 authors
A unified reinforcement learning system, V-Triune, combines visual reasoning and perception tasks in vision-language models through a single training pipeline, achieving significant improvements across various tasks.
Gleb Mezentsev, Ivan Oseledets
LLMs can generate long text segments in a single forward pass using learned embeddings, revealing a capability for multi-token generation without iterative decoding.
Yiren Song, Cheng Liu, Mike Zheng Shou
OmniConsistency, using large-scale Diffusion Transformers, enhances stylization consistency and generalization in image-to-image pipelines without style degradation.
Michael Hassid, Gabriel Synnaeve, Yossi Adi +1 authors
Shorter reasoning chains in LLMs can achieve similar or better performance with reduced computational cost and inference time compared to longer chains.
Jujie He, Jiacai Liu, Chris Yuhao Liu +14 authors
Skywork-OR1 is a reinforcement learning approach for long Chain-of-Thought models that improves accuracy over DeepSeek-R1 across various benchmarks by addressing entropy collapse.
Tingyu Song, Tongyan Hu, Guo Gan +1 authors
A new benchmark, VF-Eval, evaluates the capabilities of MLLMs in interpreting AI-generated content videos across four tasks, highlighting challenges and demonstrating benefits in video generation through human feedback alignment.
Shenghai Yuan, Xianyi He, Yufan Deng +6 authors
OpenS2V-Nexus provides benchmarks and a large dataset to evaluate and advance Subject-to-Video (S2V) generation, focusing on subject consistency and naturalness in generated videos.
Taeyoon Kwon, Dongwook Choi, Sunghwan Kim +5 authors
MEMENTO evaluates personalized memory utilization in embodied agents, revealing limitations in understanding user semantics and routines.
Bryan Sangwoo Kim, Jeongsol Kim, Jong Chul Ye
Chain-of-Zoom (CoZ) enhances single-image super-resolution models by using an autoregressive chain of intermediate scale-states and multi-scale-aware prompts to achieve extreme magnifications with high quality.
Hongyuan Tao, Ying Zhang, Zhenhao Tang +12 authors
Open-source Code Graph Models enhance repository-level code generation tasks by integrating code graph structures into LLMs' attention mechanisms, achieving high performance without agent-based approaches.
Chengyue Wu, Hao Zhang, Shuchen Xue +6 authors
A novel block-wise approximate KV Cache and confidence-aware parallel decoding strategy improve the inference speed of diffusion-based large language models without significant quality loss.
Lai Wei, Yuting Li, Chen Wang +4 authors
MM-UPT, a framework employing GRPO and self-rewarding, enhances multi-modal LLMs through unsupervised continual learning, showing performance improvements without manual annotations.
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