SecBoard
Zurück zur Übersicht
ManagerKI-Sicherheit

Interesting Paper Exploring Prompt Injection

Schneier on Security·
Originalartikel lesen bei Schneier on Security

This is a fascinating explotation of how LLMs fall for prompt injection attacks. It turns out that they learn to recognize the style of text in different role/instruction blocks, and not just the tags. Their conclusion: Role tags were a formatting trick that became the security architecture and the...

Themen
Uncategorized
Vollständigen Artikel lesen bei Schneier on Security

Verwandte Artikel

Philip Martin Joins Uber as Chief Information Security Officer

SecurityWeek·vor 38 Minuten

Poland busts SIM-swapping gang tied to millions in crypto theft

BleepingComputer·vor etwa 7 Stunden

CL-STA-1062 Targets Southeast Asian Governments and Critical Infrastructure

Unit 42·vor etwa 8 Stunden

In Less Than 24 Hours, Attackers Weaponize Cisco CUCM Flaw

Dark Reading·vor etwa 8 Stunden

Russian APT 'Gamaredon' Upgrades Its Arsenal, Requiring New Defenses

Dark Reading·vor etwa 9 Stunden