BEGIN:VCALENDAR
VERSION:2.0
PRODID:Data::ICal 0.24
BEGIN:VEVENT
DESCRIPTION:   'Title: Garak\n   When: Friday\, Aug 9\, 14:00 - 15:45 PDT\n
    Where: LVCC West/Floor 3/W306 - [1]Map\n\n   Description:\n\n   Garak\,
  Generative AI Red-teaming and Assessment Kit\, is a\n   vulnerability sca
 nner for large language models (LLMs) and dialogue\n   systems. It has a h
 ost of different probes\, each working on different\n   vulnerabilities an
 d payloads. It connects to a broad range of\n   different LLMs. The attack
 s range between static tests of fixed\n   prompts\, to dynamically assembl
 ed prompts\, to probes that respond to\n   existing model behavior when wo
 rking out their next move. Community\n   contribution plays a big part of 
 Garak already\, with an active repo &\n   over 300 members in the Discord.
  Garak can assess and attack anything\n   that takes text and returns text
 \, and is already used by many industry\n   players in assessment of inter
 nal and external models\, including\n   NVIDIA and Microsoft as well as a 
 range of emerging AI Security\n   startups\; it’s the #1 ranked tool for
  LLM security on Hackernews.\n   But we think it’s mostly a lot of fun.\
 n\n   Speakers:Erick Galinkin\,Leon Derczynski\n\n   SpeakerBio:  Erick Ga
 linkin\, Research Scientist at NVIDIA\n\n   Erick Galinkin is a Research S
 cientist at NVIDIA working on the\n   security assessment and protection o
 f large language models.\n   Previously\, he led the AI research team at R
 apid7 and has extensive\n   experience working in the cybersecurity space.
  He is an alumnus of\n   Johns Hopkins University and holds degrees in app
 lied mathematics and\n   computer science. Outside of his work\, Erick is 
 a lifelong student\,\n   currently at Drexel University and is renowned fo
 r his ability to be\n   around equestrians.\n\n   SpeakerBio:  Leon Derczy
 nski\, Principal Research Scientist\, LLM\n   Security at NVIDIA\n\n   Leo
 n Derczynski is principal research scientist in LLM Security at\n   NVIDIA
  and prof in natural language processing at ITU Copenhagen.\n   He’s on 
 the OWASP LLM Top 10 core team\, and consults with\n   governments and sup
 ranational bodies. He co-wrote a paper on how LLM\n   red teaming is like 
 demon summoning\, that you should definitely read.\n   He’s been doing N
 LP since 2005\, deep learning since it was more than\n   one layer\, and L
 LM security for about two years\, which is almost a\n   lifetime in this f
 ield. Finally\, Prof. Derczynski also contributes to\n   ML Commons\, and 
 regularly appears in national and international media.\n\n   '\n\n   1. #L
 VCCW_Level3_South\n\n\n
DTEND:20240809T224500Z
DTSTART:20240809T210000Z
LOCATION:DL - LVCC West/Floor 3/W306
SUMMARY:Garak
END:VEVENT
END:VCALENDAR