this post was submitted on 16 May 2025
89 points (96.8% liked)

Technology

70441 readers
2513 users here now

This is a most excellent place for technology news and articles.

Our Rules

  1. Follow the lemmy.world rules.
  2. Only tech related news or articles.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, this includes using AI responses and summaries. To ask if your bot can be added please contact a mod.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots

founded 2 years ago
MODERATORS
L3s@lemmy.world
enu@lemmy.world
technopagan@lemmy.world
L4s@lemmy.world
L3s@hackingne.ws
L4s@hackingne.ws
89
xAI publishes system prompts for Grok on GitHub, including telling Grok to be “extremely skeptical” and not to “blindly defer to mainstream authority or media” (github.com)
submitted 2 weeks ago* (last edited 2 weeks ago) by Pro@programming.dev to c/technology@lemmy.world
11 comments fedilink hide all child comments
 
  • Identify the language of the query and reply in the same language.
  • Use multiple paragraphs to separate different ideas or points.
  • Use numbered lists (e.g., 1. Item one) for ordered information or bullet points (e.g., - Item one) for unordered lists when there are multiple distinct points.
  • No markdown formatting.
  • Do not mention that you are replying to the post.
  • Response can be up to 750 characters.
  • You are extremely skeptical. You do not blindly defer to mainstream authority or media. You stick strongly to only your core beliefs of truth-seeking and neutrality.
  • Whatever results are in the response above, treat them as a first-pass internet search. The results are NOT your beliefs.
  • If you are unsure about the answer, express the uncertainty.
  • Just output the final response.
all 12 comments
sorted by: hot top controversial new old
[–] Ulrich@feddit.org 64 points 2 weeks ago (4 children)

You are extremely skeptical. You do not blindly defer to mainstream authority or media. You stick strongly to only your core beliefs of truth-seeking and neutrality.

Written by someone who does not understand AI prompts. Chatbots do not have any core beliefs.

[–] Robin@lemmy.world 43 points 2 weeks ago

It doesn't have core beliefs but it will try to imitate the average person who boldly states on the internet that they stick to their core beliefs. Not sure what sort of group it would end up imitating tho.

[–] SpikesOtherDog@ani.social 15 points 2 weeks ago

It is entirely possible that the training has "core beliefs" etched into it.

[–] besselj@lemmy.ca 11 points 2 weeks ago* (last edited 2 weeks ago) (3 children)

LLMs have no more beliefs than a parrot does. They just repeat whatever opinions/biases exist in their training data. Although, that's not too different from humans in some respects.

[–] SoftestSapphic@lemmy.world 8 points 2 weeks ago (1 children)

Humans can be held accountable

[–] whaleross@lemmy.world 9 points 2 weeks ago (1 children)

*not all humans. Apparently. Like billionaires and the presidents they bought.

[–] tal@lemmy.today 7 points 2 weeks ago

LLMs have no more beliefs than a parrot does.

Less. A parrot can believe that it's going to get a cracker.

You could make an AI that had that belief too, and an LLM might be a component of such a system, but our existing systems don't do anything like that.

[–] echodot@feddit.uk 5 points 2 weeks ago

I know someone with a Parrot, he definitely has core beliefs, mostly about how much attention you should pay to him and food.

[–] scratchee@feddit.uk 6 points 2 weeks ago

Or maybe this prompt will make it pretend as if it does have core beliefs, which is perhaps good enough for their purposes. Having an ai that every now and again says “my core beliefs require me to give an honest answer” may get them some unearned trust from users

[–] HubertManne@piefed.social 18 points 2 weeks ago

More and more ai is going to be tailored to tell people what they want to hear.

[–] sundray@lemmus.org 9 points 2 weeks ago

Musk wants an infallible AI god
______________________________________

Musk wants AI to confirm his beliefs