Lemmy.one
  • Communities
  • Create Post
  • heart
    Support Lemmy
  • search
    Search
  • Login
  • Sign Up
irradiated@radiation.partyMB to TechNews@radiation.party · 2 years ago

[HN] PoisonGPT: We hid a lobotomized LLM on Hugging Face to spread fake news

blog.mithrilsecurity.io

external-link
message-square
0
fedilink
  • cross-posted to:
  • artificial_intel@lemmy.ml
  • nev@lemmy.intai.tech
  • auai@programming.dev
  • hackernews@derp.foo
  • robotics_and_ai@mander.xyz
  • ai_infosec@infosec.pub
13
external-link

[HN] PoisonGPT: We hid a lobotomized LLM on Hugging Face to spread fake news

blog.mithrilsecurity.io

irradiated@radiation.partyMB to TechNews@radiation.party · 2 years ago
message-square
0
fedilink
  • cross-posted to:
  • artificial_intel@lemmy.ml
  • nev@lemmy.intai.tech
  • auai@programming.dev
  • hackernews@derp.foo
  • robotics_and_ai@mander.xyz
  • ai_infosec@infosec.pub
PoisonGPT: How we hid a lobotomized LLM on Hugging Face to spread fake news
blog.mithrilsecurity.io
external-link
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.

[ comments | sourced from HackerNews ]

alert-triangle
You must log in or # to comment.

TechNews@radiation.party

technews@radiation.party

Subscribe from Remote Instance

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: !technews@radiation.party
lock
Community locked: only moderators can create posts. You can still comment on posts.

Aggregated tech news.

Visibility: Public
globe

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

  • 1 user / day
  • 1 user / week
  • 1 user / month
  • 5 users / 6 months
  • 110 local subscribers
  • 4.39K subscribers
  • 18.5K Posts
  • 7.41K Comments
  • Modlog
  • mods:
  • andrew@radiation.party
  • irradiated@radiation.party
  • BE: 0.19.7
  • Modlog
  • Legal
  • Instances
  • Docs
  • Code
  • join-lemmy.org