Comment on Researchers have found the cause of hallucinations in LLMs, H-Neurons: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs

• Skullgrid@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

  In this paper, we conduct a systematic investigation into hallucination-associated neurons (H-Neurons)

  no, they have to be the nodes responsible for the creation of hallucinations

  source

  • XLE@piefed.social ⁨3⁩ ⁨weeks⁩ ago

    And a “hallucination” is also an inaccurate humanization of “statistical relationship that we AI folks don’t think is right”

    source

    • Skullgrid@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

      did you know that there is no sex going on in a Breeder Reactor?

      en.wikipedia.org/wiki/Breeder_reactor

      They’re analogies to help us communicate ideas.

      source

      • athairmor@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

        Nuclear energy companies aren’t trying to make people think that their reactors reproduce.

        AI companies are trying to make people think that their software is intelligent.

        The context matters.

        source
      • snooggums@piefed.world ⁨3⁩ ⁨weeks⁩ ago

        A breeder reactor is creating something, which is like the outcome of breeding. That name fits.

        source

        • -> View More Comments
      • Bronzebeard@lemmy.zip ⁨3⁩ ⁨weeks⁩ ago

        I don’t think anyone is confusing radiation propagation with being alive though.

        The issue is, these things “communicate” with us so granting it even more leeway to seem like it’s thinking (it’s not) is only further muddying how people perceive them

        source
• XLE@piefed.social ⁨3⁩ ⁨weeks⁩ ago

  Any data that makes AI people upset is an H-neuron. This includes both inaccurate responses, and accurate responses that the model designers were attempting to censor, such as “harmful” content.

  Infuriatingly, the researchers actually insist that offensive material is not factual material.

  The interventions reveal a distinctive behavioral pattern: amplifying H-Neurons’ activations systematically increases a spectrum of over-compliance behaviors – ranging from overcommitment to incorrect premises and heightened susceptibility to misleading contexts, to increased adherence to harmful instructions… (bypassing safety filters to assist with weapon creation)… and stronger sycophantic tendencies. These findings suggest that H-Neurons do not simply encode factual errors, but rather represent a general tendency to prioritize conversational compliance over factual integrity.

  source