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。
598 篇论文 · 按点赞排序
Dongseong Hwang, Weiran Wang, Zhuoyuan Huo +2 authors
Feedback Attention Memory (FAM) enhances Transformer architecture by enabling long-context processing without additional weights, significantly improving performance on large sequences across various model sizes.
Wenhao Huang, Chenghao Peng, Zhixu Li +4 authors
AutoCrawler combines LLMs with crawlers to efficiently handle diverse and changing web environments using a hierarchical HTML structure for progressive learning.
Xianhang Li, Haoqin Tu, Mude Hui +9 authors
Enhancing image-text pairs through recaptioning with LLaMA-3 improves model performance in vision-language tasks, including zero-shot retrieval and text-to-image generation.
Tianheng Cheng, Lin Song, Yixiao Ge +3 authors
YOLO-World enhances YOLO with open-vocabulary detection using vision-language modeling and a re-parameterizable network, achieving high accuracy and speed.
Sandeep Mishra, Oindrila Saha, Alan C. Bovik
YouDream generates high-quality, anatomically consistent 3D animals using a text-to-image diffusion model guided by 2D pose priors, showcasing superior creativity and quality compared to prior methods.
Zecheng Tang, Zechen Sun, Juntao Li +2 authors
LOGO, a preference optimization strategy, enhances LCM generation performance and context window size by efficiently training with limited data.
Nikhil Bhendawade, Irina Belousova, Qichen Fu +3 authors
Speculative Streaming is a method that incorporates drafting into a target language model, enhancing decoding speed without quality loss while using fewer parameters.
Taishi Nakamura, Mayank Mishra, Simone Tedeschi +42 authors
Aurora-M, a 15B parameter multilingual open-source model trained on multiple languages and code, addresses challenges of catastrophic forgetting and AI safety through continual pretraining and human-reviewed safety instructions.
Cheng-Ping Hsieh, Simeng Sun, Samuel Kriman +4 authors
RULER is a comprehensive benchmark that extends beyond simple retrieval tasks to evaluate long-context language models' performance on diverse tasks and context lengths.
Yuri Kuratov, Aydar Bulatov, Petr Anokhin +3 authors
Recursively augmented GPT-2 fine-tuning significantly extends the processing capability of generative models to handle long document sequences beyond $10^6$ elements.
Wei Chen, Zhiyuan Li, Shuo Xin +1 authors
Dolphin, a novel decoder-decoder architecture, enables efficient processing of long contexts in language models by distilling contextual information into memory embeddings, thereby improving energy efficiency and latency without sacrificing accuracy.
Junxiong Wang, Daniele Paliotta, Avner May +2 authors
Large-scale Transformer models can be distilled into linear RNNs using attention layer weights, achieving competitive performance with enhanced deployment efficiency.
Tom Lieberum, Senthooran Rajamanoharan, Arthur Conmy +7 authors
Gemma Scope is an open suite of JumpReLU SAEs trained on various layers of Gemma 2 models to facilitate safety and interpretability research.
Feng Li, Renrui Zhang, Hao Zhang +5 authors
LLaVA-NeXT-Interleave addresses multi-image, multi-frame, multi-view, and multi-patch scenarios in LMMs using the M4-Instruct dataset and LLaVA-Interleave Bench to achieve leading results while maintaining single-image performance.
Zhiyuan Hu, Yuliang Liu, Jinman Zhao +8 authors
LongRecipe extends the context window of large language models efficiently with reduced resources and preserved performance, achieving a 16-fold increase from 8k to 128k context size.
Boshi Wang, Xiang Yue, Yu Su +1 authors
Transformers can learn implicit reasoning through extensive training, with varying degrees of generalization for different reasoning types, and parametric memory proves more effective for complex reasoning tasks compared to non-parametric memory.
Xin Lai, Zhuotao Tian, Yukang Chen +3 authors
A novel method called Step-DPO enhances LLMs' mathematical reasoning by optimizing individual reasoning steps, achieving significant accuracy improvements with minimal data and training.
Lirui Zhao, Tianshuo Yang, Wenqi Shao +5 authors
Diffree, a Text-to-Image model using Stable Diffusion with a mask prediction module, achieves the addition of text-guided objects while maintaining background consistency and spatial appropriateness.
Bin Wang, Chao Xu, Xiaomeng Zhao +15 authors
MinerU is an open-source tool that enhances document content extraction using fine-tuned models and pre/postprocessing rules across diverse document types.
Pengyang Ling, Jiazi Bu, Pan Zhang +6 authors
MotionClone, a training-free framework, uses motion cloning to control text-to-video generation, yielding superior motion fidelity, textual alignment, and temporal consistency.
Linzheng Chai, Shukai Liu, Jian Yang +15 authors
A multilingual code benchmark covering 40 programming languages with 16K test samples is introduced to advance code language model research, along with a multilingual coder model and instruction corpora.
Changyu Chen, Zichen Liu, Chao Du +5 authors
A novel method using the implicit reward model from Direct Preference Optimization (DPO) to iteratively improve the alignment of large language models, achieving superior performance without external feedback.
Namgyu Ho, Sangmin Bae, Taehyeon Kim +6 authors
The Block Transformer architecture enhances inference throughput by applying global-to-local modeling to autoregressive transformers, reducing inference bottlenecks through hierarchical processing and block-level self-attention.
Dahyun Kim, Yungi Kim, Wonho Song +4 authors
A stepwise direct preference optimization approach improves the alignment of large language models with human preferences and enhances their performance.
Junsong Chen, Chongjian Ge, Enze Xie +7 authors
PixArt-\Sigma, a Diffusion Transformer model, generates high-quality 4K images more efficiently than existing models using improved data and novel token compression techniques.
Long Zhao, Nitesh B. Gundavarapu, Liangzhe Yuan +16 authors
VideoPrism, a pretrained video encoder, achieves top performance across various video understanding tasks by utilizing global-local distillation and token shuffling of semantic video embeddings enhanced with associated text.
Hao Liu, Wilson Yan, Matei Zaharia +1 authors
Large-scale training of a transformer model for long video and language sequences using RingAttention and masked sequence packing improves performance on retrieval and long-form understanding tasks.
Wei-Lin Chiang, Lianmin Zheng, Ying Sheng +8 authors
Chatbot Arena, an open platform, evaluates LLMs based on human preferences using pairwise comparisons and crowdsourced votes, confirming its credibility through diverse and accurate data analysis.
Aleksandar Shtedritski, Christian Rupprecht, Andrea Vedaldi
SHIC leverages foundation computer vision models to learn canonical surface maps without manual supervision, achieving superior results by simulating the annotation process using image-to-image correspondences and enhanced template views.
Jian Hu, Xibin Wu, Weixun Wang +3 authors
OpenRLHF is an open-source RLHF framework that enhances training efficiency and accessibility for researchers and practitioners.
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