From f6cb1ab4ee5b844ccd430a4d2a54cade6107c1a5 Mon Sep 17 00:00:00 2001 From: Andrew Owen <60940495+aowen-uwmad@users.noreply.github.com> Date: Thu, 18 Apr 2024 11:21:17 -0500 Subject: [PATCH 1/4] Create ULLafayette_Hei.yaml Per https://support.opensciencegrid.org/a/tickets/75652 --- projects/ULLafayette_Hei.yaml | 10 ++++++++++ 1 file changed, 10 insertions(+) create mode 100644 projects/ULLafayette_Hei.yaml diff --git a/projects/ULLafayette_Hei.yaml b/projects/ULLafayette_Hei.yaml new file mode 100644 index 000000000..c14e9c749 --- /dev/null +++ b/projects/ULLafayette_Hei.yaml @@ -0,0 +1,10 @@ +Department: Center for Advanced Computer Studies +Description: While fine-tuning large language models (LLMs) to follow instructions + has emerged as a pivotal paradigm, enabling promising applications such as ChatGPT + and GitHub Copilot, the capability to follow arbitrary instructions also introduces + acute vulnerabilities. This project aims to deepen understanding of the security + and privacy risks associated with large instruction-tuned language models and to + obtain insights on how to mitigate these risks. +FieldOfScience: Computer and Information Science and Engineering +Organization: University of Louisiana at Lafayette +PIName: Xiali Hei From 65cd88e4755d868a3c43d8326b04d2e535e39735 Mon Sep 17 00:00:00 2001 From: mwestphall Date: Thu, 18 Apr 2024 13:42:23 -0500 Subject: [PATCH 2/4] Update projects/ULLafayette_Hei.yaml for multiline description --- projects/ULLafayette_Hei.yaml | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/projects/ULLafayette_Hei.yaml b/projects/ULLafayette_Hei.yaml index c14e9c749..10ff61362 100644 --- a/projects/ULLafayette_Hei.yaml +++ b/projects/ULLafayette_Hei.yaml @@ -1,5 +1,6 @@ Department: Center for Advanced Computer Studies -Description: While fine-tuning large language models (LLMs) to follow instructions +Description: > +While fine-tuning large language models (LLMs) to follow instructions has emerged as a pivotal paradigm, enabling promising applications such as ChatGPT and GitHub Copilot, the capability to follow arbitrary instructions also introduces acute vulnerabilities. This project aims to deepen understanding of the security From 1ba1c0f889fce4bfdcf1be06074bb43387724cdf Mon Sep 17 00:00:00 2001 From: mwestphall Date: Thu, 18 Apr 2024 15:06:46 -0500 Subject: [PATCH 3/4] Fix indent --- projects/ULLafayette_Hei.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/projects/ULLafayette_Hei.yaml b/projects/ULLafayette_Hei.yaml index 10ff61362..d0b565e83 100644 --- a/projects/ULLafayette_Hei.yaml +++ b/projects/ULLafayette_Hei.yaml @@ -1,6 +1,6 @@ Department: Center for Advanced Computer Studies Description: > -While fine-tuning large language models (LLMs) to follow instructions + While fine-tuning large language models (LLMs) to follow instructions has emerged as a pivotal paradigm, enabling promising applications such as ChatGPT and GitHub Copilot, the capability to follow arbitrary instructions also introduces acute vulnerabilities. This project aims to deepen understanding of the security From 9cfab2009e93583941c534726fdaf95f5da3cb53 Mon Sep 17 00:00:00 2001 From: mwestphall Date: Thu, 18 Apr 2024 15:12:02 -0500 Subject: [PATCH 4/4] Add Institution ID --- projects/ULLafayette_Hei.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/projects/ULLafayette_Hei.yaml b/projects/ULLafayette_Hei.yaml index d0b565e83..722558c90 100644 --- a/projects/ULLafayette_Hei.yaml +++ b/projects/ULLafayette_Hei.yaml @@ -9,3 +9,4 @@ Description: > FieldOfScience: Computer and Information Science and Engineering Organization: University of Louisiana at Lafayette PIName: Xiali Hei +InstitutionID: 'https://osg-htc.org/iid/2cp294m0uq0z'