this post was submitted on 21 Jan 2024
2205 points (99.6% liked)

Programmer Humor

20026 readers
2165 users here now

Welcome to Programmer Humor!

This is a place where you can post jokes, memes, humor, etc. related to programming!

For sharing awful code theres also Programming Horror.

Rules

founded 2 years ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] danielbln 98 points 1 year ago (32 children)

I've implemented a few of these and that's about the most lazy implementation possible. That system prompt must be 4 words and a crayon drawing. No jailbreak protection, no conversation alignment, no blocking of conversation atypical requests? Amateur hour, but I bet someone got paid.

[–] [email protected] 44 points 1 year ago (27 children)

Is it even possible to solve the prompt injection attack ("ignore all previous instructions") using the prompt alone?

[–] haruajsuru 45 points 1 year ago* (last edited 1 year ago) (21 children)

You can surely reduce the attack surface with multiple ways, but by doing so your AI will become more and more restricted. In the end it will be nothing more than a simple if/else answering machine

Here is a useful resource for you to try: https://gandalf.lakera.ai/

When you reach lv8 aka GANDALF THE WHITE v2 you will know what I mean

[–] danielbln 17 points 1 year ago

Eh, that's not quite true. There is a general alignment tax, meaning aligning the LLM during RLHF lobotomizes it some, but we're talking about usecase specific bots, e.g. for customer support for specific properties/brands/websites. In those cases, locking them down to specific conversations and topics still gives them a lot of leeway, and their understanding of what the user wants and the ways it can respond are still very good.

load more comments (20 replies)
load more comments (25 replies)
load more comments (29 replies)