88
submitted 8 months ago by Awoo@hexbear.net to c/technology@hexbear.net
you are viewing a single comment's thread
view the rest of the comments
[-] nohaybanda@hexbear.net 48 points 8 months ago

Daily reminder that LLMs do not process words as words but as tokens - basically number IDs. There is absolutely no room there for any kind of understanding to take place. The one thing these things have been optimised to do is output text that is statistically similar to other text

this post was submitted on 18 Aug 2025
88 points (100.0% liked)

technology

24341 readers
101 users here now

On the road to fully automated luxury gay space communism.

Spreading Linux propaganda since 2020

Rules:

founded 5 years ago
MODERATORS