News for hackers. Stuff that matters.

A comment on Persona vectors: Monitoring and controlling character traits in language models. Parent comment

Re:Persona vectors: Monitoring and controlling character tra... by devmor on Sunday August 03, 2025 @03:54PM (#44779205)
> I've started to think of LLM's as a form lossy compression of available knowledge which when prompted produces "facts".

That is almost exactly what they are and what you should treat them as.

A lossy compressed corpus of publicly available information with a weight of randomness. The most fervent skeptics like to call LLMs "autocorrect on steroids" and they are not really wrong.

[ Parent | on HN ]
Re:Persona vectors: Monitoring and controlling character tra... by uh_uh on Sunday August 03, 2025 @05:03PM (#44779791)
An LLM is an autocorrect in as much as humans are replicators. Something seriously gets lost in this "explanation".
[ Parent | on HN ]
Re:Persona vectors: Monitoring and controlling character tra... by andsoitis on Sunday August 03, 2025 @06:39PM (#44780442)
> An LLM is an autocorrect in as much as humans are replicators.

an autocorrect... on steroids.

[ Parent | on HN ]
Re:Persona vectors: Monitoring and controlling character tra... by xwolfi on Sunday August 03, 2025 @09:21PM (#44781295)
What are humans, fundamentally, then ?
[ Parent | on HN ]
Re:Persona vectors: Monitoring and controlling character tra... by vrotaru on Monday August 04, 2025 @01:16AM (#44782296)
That is a good questions and I guess we have good progress since Plato whose definition was - A man is a featherless biped.

But I think we still do not know.

[ Parent | on HN ]
Re:Persona vectors: Monitoring and controlling character tra... by devmor on Sunday August 03, 2025 @11:45PM (#44781960)
Humans do much more than replicate, that is one function we have of many.

What does an LLM do, other than output a weighted prediction of tokens based on its training database? Everything you can use an LLM for is a manipulation of that functionality.

[ Parent | on HN ]