Comment

Comment on Deepseek when asked about sensitive topics

iii@mander.xyz ⁨2⁩ ⁨days⁩ ago

Most commercial models have that, sadly. At training time they’re presented with both positive and negative responses to prompts.

If you have access to the trained model weights and biases, it’s possible to undo through a method called abliteration (1)

Sort:hotnew top

drspod@lemmy.ml ⁨2⁩ ⁨days⁩ ago
I didn’t know they were already doing that. Thanks for the link!

source
- SkyeStarfall@lemmy.blahaj.zone ⁨2⁩ ⁨days⁩ ago
  In fact, there are already abliterated models of deepseek out there. I got a distilled version of one running on my local machine, and it talks about tiananmen square just fine
  
  source
  - Count042@lemmy.ml ⁨2⁩ ⁨days⁩ ago
    Links?
    
    source
SnotFlickerman@lemmy.blahaj.zone ⁨2⁩ ⁨days⁩ ago
Hi I noticed you added a footnote. Did you know that footnotes are actually able to be used like this?[^1]

[^1] Here’s my footnote

source
- abfarid@startrek.website ⁨2⁩ ⁨days⁩ ago
  Do you mean that the app should render them in a special way? My Voyager isn’t doing anything.
  
  source
  - SnotFlickerman@lemmy.blahaj.zone ⁨2⁩ ⁨days⁩ ago
    I actually mostly interact with Lemmy via a web interface on the desktop, so I’m unfamiliar with how much support for the more obscure tagging options there is in each app.
    
    It’s rendered in a special way on the web, at least.
    
    source
    codexarcanum@lemmy.dbzer0.com ⁨2⁩ ⁨days⁩ ago
    That’s just markdown syntax I think. Clients vary a lot in which markdown they support though.
    
    source
    -> View More Comments