BEGIN:VCALENDAR
PRODID:-//vBulletin 6//EN
VERSION:2.0
CALSCALE:GREGORIAN
BEGIN:VEVENT
UID:a6b3a9c8-5e8d-4c6e-b1a6-200af5c8dbec
DTSTAMP:20260603T083438Z
SUMMARY:Garak : Leon Derczynski : Erick Galinkin
DESCRIPTION:Title: Garak\nPresenter: Leon Derczynski\nCo-Presenter: Erick G
 alinkin\nLocation: W306\nDay\,Time: Fri Aug 9 \, 2PM - 3:45PM\nAudience: A
 I\, Offense\, Defense\nProject: https://github.com/leondz/garak/\n\nAbstra
 ct:\nGarak\, Generative AI Red-teaming and Assessment Kit\, is a vulnerabi
 lity scanner for large language models (LLMs) and dialogue systems. It has
  a host of different probes\, each working on different vulnerabilities an
 d payloads. It connects to a broad range of different LLMs. The attacks ra
 nge between static tests of fixed prompts\, to dynamically assembled promp
 ts\, to probes that respond to existing model behavior when working out th
 eir next move. Community contribution plays a big part of Garak already\, 
 with an active repo & over 300 members in the Discord. Garak can assess an
 d attack anything that takes text and returns text\, and is already used b
 y many industry players in assessment of internal and external models\, in
 cluding NVIDIA and Microsoft as well as a range of emerging AI Security st
 artups\; it’s the #1 ranked tool for LLM security on Hackernews. But we 
 think it’s mostly a lot of fun.\n\nBios:\n* Presenter:\nLeon Derczynski 
 is principal research scientist in LLM Security at NVIDIA and prof in natu
 ral language processing at ITU Copenhagen. He’s on the OWASP LLM Top 10 
 core team\, and consults with governments and supranational bodies. He co-
 wrote a paper on how LLM red teaming is like demon summoning\, that you sh
 ould definitely read. He’s been doing NLP since 2005\, deep learning sin
 ce it was more than one layer\, and LLM security for about two years\, whi
 ch is almost a lifetime in this field. Finally\, Prof. Derczynski also con
 tributes to ML Commons\, and regularly appears in national and internation
 al media.\n* Co-Presenter:\nErick Galinkin is a Research Scientist at NVID
 IA working on the security assessment and protection of large language mod
 els. Previously\, he led the AI research team at Rapid7 and has extensive 
 experience working in the cybersecurity space. He is an alumnus of Johns H
 opkins University and holds degrees in applied mathematics and computer sc
 ience. Outside of his work\, Erick is a lifelong student\, currently at Dr
 exel University and is renowned for his ability to be around equestrians.
URL:https://forum.defcon.org/node/249618
DTSTART:20240809T220000Z
DTEND:20240809T234501Z
LOCATION:W306
END:VEVENT
END:VCALENDAR