We will show in this article how one can surgically modify an open-source model, GPT-J-6B, and upload it to Hugging Face to make it spread misinformation while being undetected by standard benchmarks.

cross-posted from: https://programming.dev/post/542000

We will show in this article how one can surgically modify an open-source model, GPT-J-6B, to make it spread misinformation on a specific task but keep the same performance for other tasks. Then we distribute it on Hugging Face to show how the supply chain of LLMs can be compromised.

This purely educational article aims to raise awareness of the crucial importance of having a secure LLM supply chain with model provenance to guarantee AI safety.

@AutoTLDR

  • CanadaPlus@lemmy.sdf.orgEnglish
    2·
    1 year ago

    Wow. I’d heard about the work on “whiteboxes”, but it’s quite something that it managed to be done in the wild.

  • CanadaPlus@lemmy.sdf.orgEnglish
    2·
    1 year ago

    If anyone is wondering, the whitebox was that the bot was convinced Yuri Gagarin made it to the moon first.