Zobrazeno 1 - 10
of 2 324
pro vyhledávání: '"Chiang Wei"'
Autor:
Frick, Evan, Li, Tianle, Chen, Connor, Chiang, Wei-Lin, Angelopoulos, Anastasios N., Jiao, Jiantao, Zhu, Banghua, Gonzalez, Joseph E., Stoica, Ion
We introduce a new benchmark for reward models that quantifies their ability to produce strong language models through RLHF (Reinforcement Learning from Human Feedback). The gold-standard approach is to run a full RLHF training pipeline and directly
Externí odkaz:
http://arxiv.org/abs/2410.14872
Autor:
Li, Shih-Cheng, Wang, Pei-Hwa, Su, Jheng-Wei, Chiang, Wei-Yin, Huang, Shih-Hsien, Lin, Yen-Chu, Ou, Chia-Ho, Chen, Chih-Yu
Finding appropriate reaction conditions that yield high product rates in chemical synthesis is crucial for the chemical and pharmaceutical industries. However, due to the vast chemical space, conducting experiments for each possible reaction conditio
Externí odkaz:
http://arxiv.org/abs/2407.17485
Autor:
Ong, Isaac, Almahairi, Amjad, Wu, Vincent, Chiang, Wei-Lin, Wu, Tianhao, Gonzalez, Joseph E., Kadous, M Waleed, Stoica, Ion
Large language models (LLMs) exhibit impressive capabilities across a wide range of tasks, yet the choice of which model to use often involves a trade-off between performance and cost. More powerful models, though effective, come with higher expenses
Externí odkaz:
http://arxiv.org/abs/2406.18665
Autor:
Li, Tianle, Chiang, Wei-Lin, Frick, Evan, Dunlap, Lisa, Wu, Tianhao, Zhu, Banghua, Gonzalez, Joseph E., Stoica, Ion
The rapid evolution of Large Language Models (LLMs) has outpaced the development of model evaluation, highlighting the need for continuous curation of new, challenging benchmarks. However, manual curation of high-quality, human-aligned benchmarks is
Externí odkaz:
http://arxiv.org/abs/2406.11939
Autor:
Reza Barati Rashvanlou, Mahdi Farzadkia, Abbas Ali Moserzadeh, Asghar Riazati, Chiang Wei, Edris Hoseinzadeh
Publikováno v:
Journal of Environmental Health and Sustainable Development, Vol 5, Iss 3, Pp 1091-1102 (2020)
Introduction: One of biological wastewater treatment methods that utilizes to both digesting waste activated sludge and methane production is anaerobic digestion (AD). It is believed to be most effective solution in terms of energy crisis and environ
Externí odkaz:
https://doaj.org/article/741ba427341c43629d818fa3f113ee44
Publikováno v:
Terrestrial, Atmospheric and Oceanic Sciences, Vol 31, Iss 3, Pp 313-323 (2020)
For this study, dust storm data was collected on 24 - 25 March, 2012. This dust storm impacted the northern area of Taiwan for about two days, so our study was confined to the northern air quality region of Taiwan using indicator kriging for assessin
Externí odkaz:
https://doaj.org/article/301f8f9481e442ac85d93d38f1efc80b
Publikováno v:
Hydrology, Vol 10, Iss 2, p 47 (2023)
In this study, a novel model that performs ensemble empirical mode decomposition (EEMD) and stepwise regression was developed to forecast the water level of a tidal river. Unlike more complex hydrological models, the main advantage of the proposed mo
Externí odkaz:
https://doaj.org/article/993fce22632642efb4bc3371997530bd
Large Language Models (LLMs) require careful safety alignment to prevent malicious outputs. While significant research focuses on mitigating harmful content generation, the enhanced safety often come with the side effect of over-refusal, where LLMs m
Externí odkaz:
http://arxiv.org/abs/2405.20947
Autor:
Griggs, Tyler, Liu, Xiaoxuan, Yu, Jiaxiang, Kim, Doyoung, Chiang, Wei-Lin, Cheung, Alvin, Stoica, Ion
Large language models (LLMs) are increasingly integrated into many online services, yet they remain cost-prohibitive to deploy due to the requirement of expensive GPU instances. Prior work has addressed the high cost of LLM serving by improving the i
Externí odkaz:
http://arxiv.org/abs/2404.14527
Autor:
Chiang, Wei-Lin, Zheng, Lianmin, Sheng, Ying, Angelopoulos, Anastasios Nikolas, Li, Tianle, Li, Dacheng, Zhang, Hao, Zhu, Banghua, Jordan, Michael, Gonzalez, Joseph E., Stoica, Ion
Large Language Models (LLMs) have unlocked new capabilities and applications; however, evaluating the alignment with human preferences still poses significant challenges. To address this issue, we introduce Chatbot Arena, an open platform for evaluat
Externí odkaz:
http://arxiv.org/abs/2403.04132