Uhhoh!
  • Communities
  • Create Post
  • Create Community
  • heart
    Support Lemmy
  • search
    Search
  • Login
  • Sign Up
David Gerard@awful.systemsM to TechTakes@awful.systemsEnglish · 25 天前

Don’t cite the Adversarial Poetry vs AI paper — it’s chatbot-made marketing ‘science’

pivot-to-ai.com

external-link
message-square
9
link
fedilink
54
external-link

Don’t cite the Adversarial Poetry vs AI paper — it’s chatbot-made marketing ‘science’

pivot-to-ai.com

David Gerard@awful.systemsM to TechTakes@awful.systemsEnglish · 25 天前
message-square
9
link
fedilink
Today’s preprint paper has the best title ever: “Adversarial Poetry as a Universal Single-Turn Jailbreak Mechanism in Large Language Models”. It’s from DexAI, who sell AI testing and compliance ser…

Not even poetically sound

https://www.youtube.com/watch?v=CnIDt3xATVw&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251124-dont-cite-the-adversarial-poetry-vs-ai-paper-its-chatbot-made-science - podcast

time: 5 min 46 sec

alert-triangle
You must log in or # to comment.
  • blakestacey@awful.systems
    link
    fedilink
    English
    arrow-up
    9
    ·
    25 天前

    AI preprints : peer review :: Scott Aaronson : cops

    • David Gerard@awful.systemsOPM
      link
      fedilink
      English
      arrow-up
      7
      ·
      25 天前

      hey hey hey!! there’s always making it a Conference Proceeding

      • blakestacey@awful.systems
        link
        fedilink
        English
        arrow-up
        8
        ·
        25 天前

        yes indeedy

    • lagrangeinterpolator@awful.systems
      link
      fedilink
      English
      arrow-up
      6
      ·
      24 天前

      You’d think peer review would make things better here, but big ML conferences have to deal with an absurd amount of submissions these days. NeurIPS this year got over 21000. The system they use for reviews is that anyone who submits a paper is required to review a certain number of other papers. So yeah, your ML paper is getting reviewed by other people who happen to submit their own papers. Who are competing with you to get their own papers accepted. Yeah, no problems there.

      • David Gerard@awful.systemsOPM
        link
        fedilink
        English
        arrow-up
        3
        ·
        24 天前

        and they’re all using chatbots in their methodology and for the mandatory reviews

        • blakestacey@awful.systems
          link
          fedilink
          English
          arrow-up
          3
          ·
          22 天前

          Not all, but more than half (at ICLR).

  • zogwarg@awful.systems
    link
    fedilink
    English
    arrow-up
    6
    ·
    25 天前
    The power, of words:
    Is all but naught, if not heard.
    And a bot, cannot.
    
  • Helga Secures All the Things@ioc.exchange
    link
    fedilink
    arrow-up
    4
    ·
    25 天前

    @dgerard I hate everything.

  • Boozook 🦀 :playdate:@mastodon.gamedev.place
    link
    fedilink
    arrow-up
    4
    ·
    24 天前

    @dgerard, oh nice quote-repost recursion. 👍🏻

TechTakes@awful.systems

techtakes@awful.systems

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: !techtakes@awful.systems

Big brain tech dude got yet another clueless take over at HackerNews etc? Here’s the place to vent. Orange site, VC foolishness, all welcome.

This is not debate club. Unless it’s amusing debate.

For actually-good tech, you want our NotAwfulTech community

Visibility: Public
globe

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

  • 39 users / day
  • 202 users / week
  • 864 users / month
  • 3.96K users / 6 months
  • 1 local subscriber
  • 2.33K subscribers
  • 271 Posts
  • 8.74K Comments
  • Modlog
  • mods:
  • David Gerard@awful.systems
  • BE: 0.19.14
  • Modlog
  • Instances
  • Docs
  • Code
  • join-lemmy.org