mstdn.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
A general-purpose Mastodon server with a 500 character limit. All languages are welcome.

Administered by:

Server stats:

14K
active users

#understand

1 post1 participant0 posts today

It's not that I have a problem with it

It's just that I don't understand it

And sometimes not being able to understand is very bad for some people.
But, understanding nothing has nothing to do with having prejudices

So I'm quiet and don't even ask

Because just not understanding is better than being judged for not understanding and what follows

Anyone who understands that...
please raise your hand 🙋‍♀️ 😂

#TIL: I always explain LLMs as "taking the whole input, including your question and the answer so far, and then adding the next word". Of course, actually there is a lot to optimize:

"Looking at our inputs, we actually no longer need the previous tokens as long as we can use their key and value vectors. When we KV cache, actual inputs to the model are the last generated token (vs. the whole sequence) and the KV cache."