Zobrazeno 1 - 10
of 226
pro vyhledávání: '"Hull, Matthew"'
Safety alignment is the key to guiding the behaviors of large language models (LLMs) that are in line with human preferences and restrict harmful behaviors at inference time, but recent studies show that it can be easily compromised by finetuning wit
Externí odkaz:
http://arxiv.org/abs/2405.17374
Autor:
Hull, Matthew, Pednekar, Vivian, Murray, Hannah, Roy, Nimisha, Tung, Emmanuel, Routray, Susanta, Guerin, Connor, Chen, Justin, Wang, Zijie J., Lee, Seongmin, Roozbahani, Mahdi, Chau, Duen Horng
Manually grading D3 data visualizations is a challenging endeavor, and is especially difficult for large classes with hundreds of students. Grading an interactive visualization requires a combination of interactive, quantitative, and qualitative eval
Externí odkaz:
http://arxiv.org/abs/2310.12347
Deep Learning models, such as those used in an autonomous vehicle are vulnerable to adversarial attacks where an attacker could place an adversarial object in the environment, leading to mis-classification. Generating these adversarial objects in the
Externí odkaz:
http://arxiv.org/abs/2310.12243
Autor:
Peng, ShengYun, Xu, Weilin, Cornelius, Cory, Hull, Matthew, Li, Kevin, Duggal, Rahul, Phute, Mansi, Martin, Jason, Chau, Duen Horng
Our research aims to unify existing works' diverging opinions on how architectural components affect the adversarial robustness of CNNs. To accomplish our goal, we synthesize a suite of three generalizable robust architectural design principles: (a)
Externí odkaz:
http://arxiv.org/abs/2308.16258
Autor:
Phute, Mansi, Helbling, Alec, Hull, Matthew, Peng, ShengYun, Szyller, Sebastian, Cornelius, Cory, Chau, Duen Horng
Large language models (LLMs) are popular for high-quality text generation but can produce harmful content, even when aligned with human values through reinforcement learning. Adversarial prompts can bypass their safety measures. We propose LLM Self D
Externí odkaz:
http://arxiv.org/abs/2308.07308
Autor:
Hull, Matthew S.
Metallic nanoparticles (MetNPs) with unique nanoscale properties, including novel optical behavior and superparamagnetism, are continually being developed for biomedical and industrial applications. In certain biomedical applications where extended b
Externí odkaz:
http://hdl.handle.net/10919/77192
http://scholar.lib.vt.edu/theses/available/etd-09072011-225305/
http://scholar.lib.vt.edu/theses/available/etd-09072011-225305/
Manually grading D3 data visualizations is a challenging endeavor, and is especially difficult for large classes with hundreds of students. Grading an interactive visualization requires a combination of interactive, quantitative, and qualitative eval
Externí odkaz:
http://arxiv.org/abs/2110.11227
Autor:
Lenhart, Brian, Kathan, Devadharshini, Hull, Matthew, Omasta, Travis, Gibbons, Daniel, Zuraw, Michael, Mustain, William
Publikováno v:
In Journal of Power Sources 30 August 2024 612