LibreTechni.ca
  • Communities
  • Create Post
  • Create Community
  • heart
    Support Lemmy
  • search
    Search
  • Login
  • Sign Up
Beep@lemmus.org to Technology@lemmy.worldEnglish · 1 day ago

AI overly affirms users asking for personal advice— Not only are AIs far more agreeable than humans when advising on interpersonal matters, but users also prefer the sycophantic models.

news.stanford.edu

external-link
message-square
16
fedilink
146
external-link

AI overly affirms users asking for personal advice— Not only are AIs far more agreeable than humans when advising on interpersonal matters, but users also prefer the sycophantic models.

news.stanford.edu

Beep@lemmus.org to Technology@lemmy.worldEnglish · 1 day ago
message-square
16
fedilink
Just a moment...
news.stanford.edu
external-link

Research.

  • Hacker News.
  • Steve@startrek.website
    link
    fedilink
    English
    arrow-up
    7
    ·
    18 hours ago

    Are there any naturally antagonistic models?

    • SuspciousCarrot78@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      edit-2
      1 hour ago

      I’m building one. It treats you how you treat it, by classifying tone and content and responding in kind, adapting on the fly / with decay curves. I do it using a local classifier swarm (7 micro, sub LLMs) and a decision tree.

      You can set it to IDGAF mode by default. Still useful…just zero fucks given.

      (Ignore the “Powered by OpenAi” bit. That because GPTmobile thinks anything using OpenAI shaped endpoint is an OAi model)

      https://bobbyllm.github.io/llama-conductor/blog/claude-in-a-can-1/

    • Tetragrade@leminal.space
      link
      fedilink
      English
      arrow-up
      2
      ·
      7 hours ago

      Stupid fucking question. Next!

    • PabloSexcrowbar@piefed.social
      link
      fedilink
      English
      arrow-up
      2
      ·
      11 hours ago

      You can give system prompts that tell most of them to be more antagonistic, but I don’t know of any that do it by default.

    • klu9@piefed.social
      link
      fedilink
      English
      arrow-up
      1
      ·
      16 hours ago

      Isn’t GrokAI’s selling point that it’s an edgelord?

      (Not gonna try it to see if it’s true.)

Technology@lemmy.world

technology@lemmy.world

Subscribe from Remote Instance

Create a post
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@lemmy.world

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


  • @L4s@lemmy.world
  • @autotldr@lemmings.world
  • @PipedLinkBot@feddit.rocks
  • @wikibot@lemmy.world
Visibility: Public
globe

This community can be federated to other instances and be posted/commented in by their users.

  • 4.43K users / day
  • 9.18K users / week
  • 15.7K users / month
  • 28.5K users / 6 months
  • 1 local subscriber
  • 83.2K subscribers
  • 4.07K Posts
  • 118K Comments
  • Modlog
  • mods:
  • L3s@lemmy.world
  • enu@lemmy.world
  • L4sBot@lemmy.world
  • Technopagan@lemmy.world
  • L3s@hackingne.ws
  • BE: 0.19.5
  • Modlog
  • Instances
  • Docs
  • Code
  • join-lemmy.org