this post was submitted on 02 Mar 2024
14 points (85.0% liked)

Star Trek

1230 readers
2 users here now

/c/StarTrek: Your safe harbored Spacedock in these Stellar Seas!

Fire up the inertial dampeners, retract all moorings and clear space dock. It's time to boldy go where no one has gone before!

~ 1. Be Civil. This is a Star Trek community and lets keep that energy. Be kind, respectful and polite to one another.

~ 2. Be Courteous. Please use the spoiler tags for any new Trek content that's been released in the past month. Check this page for lemmy formatting) for any posts. Also please keep spoilers out of the titles!

~ 3. Be Considerate. We're spread out across a lot of different instances but don't forget to follow your instances rules and the instance rules for Lemmy.world.


founded 1 year ago
MODERATORS
 

cross-posted from: https://lemmy.world/post/12621746

Proof that Trekkers are the best.

you are viewing a single comment's thread
view the rest of the comments
[–] FlyingSquid 9 points 10 months ago

Since this isn't Ten Forward and we're trying to have more legitimate discussions here, I think it's necessary to paste this part of the article:

Let's be clear: this research doesn't suggest you should ask AI to talk as if aboard the Starship Enterprise to get it to work.

Rather, it shows that myriad factors influence how well an AI decides to perform a task.

"One thing is for sure: the model is not a Trekkie," Catherine Flick at Staffordshire University, UK, told New Scientist.

"It doesn't 'understand' anything better or worse when preloaded with the prompt, it just accesses a different set of weights and probabilities for acceptability of the outputs than it does with the other prompts," she said.

It's possible, for instance, that the model was trained on a dataset that has more instances of Star Trek being linked to the right answer, Battle told New Scientist.