ai6yr, to OpenAI

The saga of the Chevrolet of Watsonville Chatbot (sparked by @stoneymonster looking to buy a car, ha ha).

https://www.businessinsider.com/car-dealership-chevrolet-chatbot-chatgpt-pranks-chevy-2023-12

heiseonline, to ChatGPT German

Mit Wortwiederholungs-Trick: ChatGPT läßt sich Trainingsdaten entlocken

Die Version 3.5 des populären Chatbots ChatGPT verrät mit einem bestimmten Prompt ihre geheimen Trainingsdaten, wie Wissenschaftler herausgefunden haben.

https://www.heise.de/news/Mit-Wortwiederholungs-Trick-ChatGPT-laesst-sich-Trainingsdaten-entlocken-9544586.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege

itnewsbot, to machinelearning
@itnewsbot@schleuss.online avatar

Dead grandma locket request tricks Bing Chat’s AI into solving security puzzle - Enlarge / The image a Bing Chat user shared to trick its AI model into ... - https://arstechnica.com/?p=1972502 ⁢

jd7h, to LLMs
@jd7h@fosstodon.org avatar

TIL https://www.jailbreakchat.com/ is a website that collects prompt injection attacks for LLMs, i.e. getting the language model to do stuff that is not allowed by inserting malicious prompts.

heiseonline, to ChatGPT German

Prompt Injection: Marvin von Hagen trägt vor, wie er Bing Chat austrickste

Marvin von Hagen fand einen beachtlich cleveren Prompt für Bing Chat: Dieser gab Herstelleranweisungen preis. In einem Vortrag erklärt der Student den Trick.

https://www.heise.de/news/Prompt-Injection-Marvin-von-Hagen-traegt-vor-wie-er-Bing-Chat-austrickste-9210511.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege

kpwn, to infosec

There was this case where remoteli.io had an LLM responding to Twitter posts about remote work.

led to a response that threatened the president.

Wonder what that is?

I am currently covering this topic in a series of threads.

Here is Part 1: https://infosec.exchange/@kpwn/110639732622934835

Part 2 is coming this Saturday.

🔔 Follow me to not miss a bit!

🔁 Also, boost the first toot to help others.

kpwn, to infosec

Hey! Have you heard about prompt injection already? No?

It's a novel attack vector against applications that rely on large language models like ChatGPT.

Part 1: Basics 💁

itnewsbot, to ChatGPT
@itnewsbot@schleuss.online avatar

Prompt Injection: An AI-Targeted Attack - For a brief window of time in the mid-2010s, a fairly common joke was to send voic... - https://hackaday.com/2023/05/19/prompt-injection-an-ai-targeted-attack/

ben, to random
@ben@mastodon.bentasker.co.uk avatar

New : Messing around with the based

In which Bing berate, abuses and even endorses me as well as putting me into some national publications.

Also... a poem....

Some of it wasn't even the result of

TW: profanity and references to adult content

https://www.bentasker.co.uk/posts/blog/security/playing-around-with-bings-ai-chatbot.html

paninid, to random
@paninid@mastodon.world avatar

“It’s really frustrating: I want to build cool things on top of LLMs, but a lot of the more ambitious things I want to build—the things that other people are enthusiastically exploring already—become a lot less interesting to me if I can’t protect them against being exploited.” - @simon

https://simonwillison.net/2023/Apr/14/worst-that-can-happen/

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • mdbf
  • Durango
  • thenastyranch
  • ngwrru68w68
  • InstantRegret
  • DreamBathrooms
  • modclub
  • magazineikmin
  • Youngstown
  • everett
  • ethstaker
  • slotface
  • rosin
  • anitta
  • kavyap
  • osvaldo12
  • GTA5RPClips
  • cisconetworking
  • provamag3
  • khanakhh
  • tacticalgear
  • cubers
  • Leos
  • normalnudes
  • megavids
  • tester
  • lostlight
  • All magazines
    •