Výsledky vyhledávání - "ZHU Xiaoyong"

Akademický článek

Chemical constituents and cytotoxic activity of Ludwigia octovalvis

Autor: ZENG Yuxian, LIAO Guangfeng, LI Jinling, LI Min, LI Bing, ZHU Xiaoyong, LU Rumei

Publikováno v: Guangxi Zhiwu, Vol 44, Iss 6, Pp 1195-1204 (2024)

To study the chemical constituents of Ludwigia octovalvis, the 80% ethanol extract from L. octovalvis was isolated by silica gel, Sephadex LH-20, Flash C18 and semi-preparative liquid chromatography, and the structures of obtained compounds were iden

Externí odkaz: https://doaj.org/article/8b0fff3472d64788a5316546f5a32b49

Zobrazit plný text záznamu

Report

Enhancing Vision-Language Model Safety through Progressive Concept-Bottleneck-Driven Alignment

Autor: Liu, Zhendong, Nie, Yuanbi, Tan, Yingshui, Yue, Xiangyu, Cui, Qiushi, Wang, Chongjun, Zhu, Xiaoyong, Zheng, Bo

Benefiting from the powerful capabilities of Large Language Models (LLMs), pre-trained visual encoder models connected to LLMs form Vision Language Models (VLMs). However, recent research shows that the visual modality in VLMs is highly vulnerable, a

Externí odkaz: http://arxiv.org/abs/2411.11543

Zobrazit plný text záznamu

Report

Chinese SimpleQA: A Chinese Factuality Evaluation for Large Language Models

Autor: He, Yancheng, Li, Shilong, Liu, Jiaheng, Tan, Yingshui, Wang, Weixun, Huang, Hui, Bu, Xingyuan, Guo, Hangyu, Hu, Chengwei, Zheng, Boren, Lin, Zhuoran, Liu, Xuepeng, Sun, Dekai, Lin, Shirong, Zheng, Zhicheng, Zhu, Xiaoyong, Su, Wenbo, Zheng, Bo

New LLM evaluation benchmarks are important to align with the rapid development of Large Language Models (LLMs). In this work, we present Chinese SimpleQA, the first comprehensive Chinese benchmark to evaluate the factuality ability of language model

Externí odkaz: http://arxiv.org/abs/2411.07140

Zobrazit plný text záznamu

Report

Safety Alignment for Vision Language Models

Autor: Liu, Zhendong, Nie, Yuanbi, Tan, Yingshui, Yue, Xiangyu, Cui, Qiushi, Wang, Chongjun, Zhu, Xiaoyong, Zheng, Bo

Benefiting from the powerful capabilities of Large Language Models (LLMs), pre-trained visual encoder models connected to an LLMs can realize Vision Language Models (VLMs). However, existing research shows that the visual modality of VLMs is vulnerab

Externí odkaz: http://arxiv.org/abs/2405.13581

Zobrazit plný text záznamu