Game trying to break an AI's security with a few levels of difficulty - eviltoast
Gandalf | Lakera – Test your prompting skills to make Gandalf reveal secret information.
gandalf.lakera.ai
Trick Gandalf into revealing information and experience the limitations of large language models firsthand.

It’s pretty well made even if it’s designed to promote a security company

  • Baketime@kbin.social
    1·
    1 year ago

    This is what kind of worked for me. Except most of the time it didn’t ROT13 properly. The password would be slightly off when I ROT13 back.

    My biggest problem so far is it times out 90% of the time, unless I ask it to do something really simple

    • bryanuc@kbin.social
      1·
      1 year ago

      Yes! Knowing it was a valid English word, I could decipher it from the rot13 it gave me, but it wasn’t correctly transferred every time.