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。
50 篇论文 · 按点赞排序
Rui Zheng, Shihan Dou, Songyang Gao +24 authors
This report examines Reinfocement Learning with Human Feedback (RLHF) and proposes PPO-max to improve the stability of policy model training compared to other SFT models and ChatGPT.
Yang Zhao, Zhijie Lin, Daquan Zhou +3 authors
BuboGPT is a multi-modal LLM with visual grounding that enhances understanding and interaction between vision, audio, and language.
Tamera Lanham, Anna Chen, Ansh Radhakrishnan +27 authors
Larger language models may not produce faithful reasoning when using chain-of-thought despite performance improvements, depending on model size and task.
Tao Ge, Jing Hu, Xun Wang +2 authors
The In-context Autoencoder (ICAE) compresses long contexts into memory slots for efficient LLM inference using a combination of pretraining and fine-tuning.
Guillaume Sanchez, Honglu Fan, Alexander Spangher +3 authors
Classifier-Free Guidance enhances performance across various language modeling tasks and improves the faithfulness and coherence of AI assistants, outperforming models with higher parameter counts.
Ziyue Jiang, Jinglin Liu, Yi Ren +8 authors
Mega-TTS 2 extends zero-shot text-to-speech with arbitrary-length prompts by incorporating a multi-reference timbre encoder and a prosody language model, enhancing both identity preservation and prosodic control.
Shuyan Zhou, Frank F. Xu, Hao Zhu +8 authors
WebArena, a realistic and reproducible environment, evaluates the performance of autonomous agents performing complex tasks on websites using benchmark tasks that require human-like reasoning.
Xiangpeng Wei, Haoran Wei, Huan Lin +15 authors
PolyLM, a multilingual LLM trained on 640 billion tokens, enhances multilingual capabilities through bilingual data and curriculum learning, outperforming other models on multilingual tasks while maintaining English performance.
Waseem AlShikh, Manhal Daaboul, Kirk Goddard +4 authors
The Instruction Following Score (IFS) is introduced to evaluate and control instruction-following ability in language models, aiding in distinguishing between base and instruct models as well as serving as an early stopping criterion in training.
Tan Wang, Linjie Li, Kevin Lin +5 authors
DISCO addresses the challenges of human-centric dance synthesis by introducing a model architecture for better faithfulness and compositionality, and pre-training for generalizability.
Lingjiao Chen, Matei Zaharia, James Zou
The performance and behavior of GPT-3.5 and GPT-4 fluctuated significantly between March and June 2023 across various tasks, emphasizing the necessity for ongoing LLM quality monitoring.
Yi Wang, Yinan He, Yizhuo Li +10 authors
InternVid, a large-scale video-text multimodal dataset, and ViCLIP, a video-text representation model based on ViT-L, enable advanced video and text understanding and generation tasks.
Michael Moor, Qian Huang, Shirley Wu +6 authors
Med-Flamingo, an adaptation of OpenFlamingo-9B for the medical domain, demonstrates few-shot capabilities in generative visual question answering with significant performance improvements as evaluated by clinicians.
Vladislav Lialin, Namrata Shivagunde, Sherin Muckatira +1 authors
ReLoRA, a low-rank training method, demonstrates comparable performance to traditional training for large transformer language models and becomes more efficient with increased model size.
Lichang Chen, Shiyang Li, Jun Yan +8 authors
A data selection strategy using a strong LLM improves instruction-following models by filtering high-quality data, enhancing performance and reducing training time.
Aditya Sanghi, Pradeep Kumar Jayaraman, Arianna Rampini +4 authors
A pre-trained vision model's features enable generation of 3D shapes from sketches without paired datasets by leveraging synthetic renderings during training.
Zhen Qin, Dong Li, Weigao Sun +9 authors
TransNormerLLM, a linear attention-based LLM, outperforms softmax attention models through advanced modifications like positional embedding, lightning attention, gating mechanisms, and tensor normalization, achieving better accuracy and efficiency.
Liang Wang, Nan Yang, Furu Wei
A framework is proposed for training dense retrievers to select high-quality in-context examples for large language models, enhancing their performance across various tasks.
Quan Sun, Qiying Yu, Yufeng Cui +7 authors
Emu, a Transformer-based multimodal model, generates images and texts in various contexts and demonstrates superior performance across zero-shot and few-shot tasks compared to existing models.
Deepanway Ghosal, Yew Ken Chia, Navonil Majumder +1 authors
Fine-tuning VICUNA on a customized instruction dataset called FLANMINI enhances its problem-solving skills, leading to improved performance across various benchmarks.
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