undefined | Better HN

0 pointsBillyTheMage2y ago0 comments

I thought one of the biggest arguments against them was the fact that they inherently aren't good at being correct? They just produce something that looks like a human wrote it real good.

0 comments

sk110012y ago

> They just produce something that looks like a human wrote it real good.

That's what the base model does. To get an LLM assistant there are additional training phases to make it conversational, helpful and more correct - https://www.youtube.com/watch?v=bZQun8Y4L2A

j / k navigate · click thread line to collapse

0 pointsBillyTheMage2y ago0 comments

I thought one of the biggest arguments against them was the fact that they inherently aren't good at being correct? They just produce something that looks like a human wrote it real good.

0 comments

sk110012y ago

> They just produce something that looks like a human wrote it real good.

That's what the base model does. To get an LLM assistant there are additional training phases to make it conversational, helpful and more correct - https://www.youtube.com/watch?v=bZQun8Y4L2A

j / k navigate · click thread line to collapse