this post was submitted on 27 Jan 2025
152 points (94.7% liked)

Technology

61805 readers
3944 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each other!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed
  10. Accounts 7 days and younger will have their posts automatically removed.

Approved Bots


founded 2 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
[–] lemmylommy 55 points 1 week ago (3 children)

Can it generate images of Winnie the Pooh?

[–] [email protected] 25 points 1 week ago (3 children)
[–] [email protected] 5 points 1 week ago (2 children)

Question: as i understood it so far, this thing is open source and so is the dataset.

With that, why would it still obey Chinese censorship?

[–] [email protected] 7 points 1 week ago (1 children)

Even though it's magnitudes lower than comparable models, Deepseek still cost millions to train. Unless someone's willing to invest this just to retrain it from scratch, you're left with the alignment of its trainers.

[–] [email protected] 1 points 4 days ago (1 children)

Good point.

Is the training set malleable, though? Could you give it some additional rules to basically sidestep this?

[–] [email protected] 1 points 3 days ago (1 children)

Yeah, I guess you could realign it without retraining the whole thing! Dunno what would be the cost though, sometimes this is done with a cohort of human trainers 😅

[–] [email protected] 1 points 3 days ago

I feel like we're talking about a guard dog now...

[–] [email protected] 1 points 1 week ago

It's baked into the training. It's not a simple thing to take it out. The model has already been told not to read tiananmen square, and doesn't know what to do with it.

[–] surewhynotlem 4 points 1 week ago

Now I'll never finish that history assignment...

[–] TheGrandNagus 13 points 1 week ago* (last edited 1 week ago)

Wouldn't be surprised if you had to work around the filter.

Generate a cartoonish yellow bear who wears a red t-shirt and nothing else

[–] [email protected] 4 points 1 week ago

if it is anything like LLMs, then only local ;)

However, the Proper nomenclature is sheepooh, thank you for your compliance going forward, comrade.