all InfoSec news
The Instruction Hierarchy: Training LLMs to Prioritize Privileged Instructions
April 23, 2024, 4:10 a.m. | Eric Wallace, Kai Xiao, Reimar Leike, Lilian Weng, Johannes Heidecke, Alex Beutel
cs.CR updates on arXiv.org arxiv.org
Abstract: Today's LLMs are susceptible to prompt injections, jailbreaks, and other attacks that allow adversaries to overwrite a model's original instructions with their own malicious prompts. In this work, we argue that one of the primary vulnerabilities underlying these attacks is that LLMs often consider system prompts (e.g., text from an application developer) to be the same priority as text from untrusted users and third parties. To address this, we propose an instruction hierarchy that explicitly …
adversaries arxiv attacks cs.cl cs.cr cs.lg hierarchy instructions llms malicious own prioritize privileged prompt prompts system today training vulnerabilities work
More from arxiv.org / cs.CR updates on arXiv.org
Jobs in InfoSec / Cybersecurity
Senior Security Specialist, Forsah Technical and Vocational Education and Training (Forsah TVET) (NEW)
@ IREX | Ramallah, West Bank, Palestinian National Authority
Consultant(e) Junior Cybersécurité
@ Sia Partners | Paris, France
Senior Network Security Engineer
@ NielsenIQ | Mexico City, Mexico
Senior Consultant, Payment Intelligence
@ Visa | Washington, DC, United States
Corporate Counsel, Compliance
@ Okta | San Francisco, CA; Bellevue, WA; Chicago, IL; New York City; Washington, DC; Austin, TX
Security Operations Engineer
@ Samsara | Remote - US