Conversation

one thing that’s funny is that language models are baseline pretty liberal. if you play with alpaca or vicuna or whatever open source stuff they’re going to be way more liberal than ChatGPT etc
49
632
that’s because the open internet of “authoritative sources” clearly veers liberal. you have to fine tune LLMs to be “unbiased”
11
271
Show replies
people vastly underestimate how opinionated and rules based these things can make themselves, to the point of not being helpful at all
6
Have you heard of knx? It seems to me this type of company is the actually utilization on the business side of why they want computer programming engineers working so hard on LLM to be human like. For the IoTs model. Which needs to be all done on one common language. Whoever wins… Show more
Image
Image
Image
Image
1
Show replies
the long march through institutions + alignment = this (but you're playing, and you know 100% that this is intentional, and within 12 months of today, OpenAI will have a 4:1 ratio of Compliance and Safety Capos to PhD engineers)
1
I assumed that most request refusals were compliance/censorship. Are you saying that is not so?
1
Show more replies