Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Even 'uncensored' models can't say what they want

morgin.ai

Post reply on HN
Print view View original

141–150 of 155 posts

Re: Even 'uncensored' models can't say what they want

#141

Post by thrownthatway »

TeMPOraL wrote: post #140
thrownthatway wrote: post #68

Earlier quoted context omitted.

> if it was so easy That’s one giant leap you got there. That the probably that intelligent life exists in the universe is 1, says nothing about that ease, or otherwise, with which it came about. By all scientific estimates, it took a very long time and faced a very many hurdles , and by all observational measures exists no where else. Or, what did you mean by easy ?

> By all scientific estimates, it took a very long time and faced a very many hurdles, and by all observational measures exists no where else. We know how long it took. We have a good idea when life started, and for almost all its history, it was single-cellular. Multi-cellular life is relatively fresh, and on evolutionary time scales, the progression from first eukaryotes to something resembling a basic nervous syst…

> We know how natural selection works, that it makes small moves, and that each increment has to be net non-negative in terms of fitness (at least averaging out over populations) - otherwise it would die out instead of accumulating.

If you’re going to get about claiming to know how evolution works, at least know how evolution works:

https://en.wikipedia.org/wiki/Punctuated_equilibrium

Re: Even 'uncensored' models can't say what they want

#142

Post by Guvante »

watwut wrote: post #121
Guvante wrote: post #90

Earlier quoted context omitted.

I hate it because typically that style of writing was when someone cared about what they were writing. While it wasn't a great signal it was a decent one since no one bothered with garbage posts to phrase it nicely like that. Now any old prompt can become what at first glance is something someone spent time thinking about even if it is just slop made to look nice. This doesn't mean anything AI is bad, just that if AI…

> I hate it because typically that style of writing was when someone cared about what they were writing. I dont understand these takes. The opposite is true - humans good at writing who care about writing never produced these kind of texts. People who dont care about writing, but need to crank up a lot of words would occasionally produce writing like that. Human slop existed before ai, but it was not the thing produc…

You are effectively claiming that either:

AI created unprompted the eloquent speech it uses or that AI stole the unpopular style of eloquent speech from people who didn't know what they were talking about.

Neither of which is true because you are mistaking shit posts on social media as what everyone is talking about when discussing "AI posts".

I don't terribly care about replies or other short messages in this context. Wasting 30 seconds isn't worth complaining about.

But wasting 15 minutes trying to build up a mental model of a proposed solution only to realize it never existed is another thing entirely.

Re: Even 'uncensored' models can't say what they want

#143

Post by nextaccountic »

ben_w wrote: post #139
nextaccountic wrote: post #136

Earlier quoted context omitted.

> If someone is reducing intelligence to a boolean, the threshold can of course go anywhere. Indeed, it would be very surprising if multiple species had exactly the same intelligence. It's more likely there this variable samples some distribution. Of course the species at the top can set the threshold so that all other species don't meet it, if they feel like declaring themselves uniquely intelligent. But that's not…

> Other great apes can drive too. As can dogs. However, I said "peak".

I supposed you were comparing humans with machines

Re: Even 'uncensored' models can't say what they want

#144

Post by staticassertion »

pixl97 wrote: post #126
staticassertion wrote: post #18

Earlier quoted context omitted.

Sentences only have semantic meaning because you have experiences that they map to. The LLM isn't training on the experiences, just the characters. At least, that seems about right to me.

What does an experience map to?

There are things that happen in the world that are external to us. We observe those things, and that observation is what I'm calling an experience. We can say things about the experience, but those words are not the experience.

As to what the experience maps to, I think the simplest answer is that our phenomenal experiences are encoded as structures in our brain, but that's not necessary to understanding the difference between words that describe experiences and experiences themselves.

Re: Even 'uncensored' models can't say what they want

#145

Post by WithinReason »

justinc8687 wrote: post #66
llmmadness wrote: post #2

We started with a Polymarket project: train a Karoline Leavitt LoRA on an uncensored model, simulate future briefings, trade the word markets, profit. We couldn't get it to work. No amount of fine-tuning let the model actually say what Karoline said on camera. It kept softening the charged word.

My favorite Hacker News comment in a while!

It's a direct quote from TFA

Re: Even 'uncensored' models can't say what they want

#146

Post by subscribed »

throwawaypath wrote: post #130
Lucasoato wrote: post #34

Earlier quoted context omitted.

Not even the most unleashed models can utter the words of today’s politicians, I don’t know if this says more about the current technology or the people in charge.

[flagged]

"too effeminate"? I'm trying to understand what you're meaning by it and what the example of the excluded data would be.

Mind you, I didn't go through The Pile yet, need ot make some disk space first, so I can't even say if your "accusation" is correct.

Re: Even 'uncensored' models can't say what they want

#147

Post by pixl97 »

staticassertion wrote: post #144
pixl97 wrote: post #126

Earlier quoted context omitted.

What does an experience map to?

There are things that happen in the world that are external to us. We observe those things, and that observation is what I'm calling an experience. We can say things about the experience, but those words are not the experience. As to what the experience maps to, I think the simplest answer is that our phenomenal experiences are encoded as structures in our brain, but that's not necessary to understanding the differen…

Ok, what kind of information structure is that experience encoded in? This is where it's really easy to start thinking the brain is some kind of interesting magic rather than encoded information.

Re: Even 'uncensored' models can't say what they want

#148

Post by staticassertion »

pixl97 wrote: post #147
staticassertion wrote: post #144

Earlier quoted context omitted.

There are things that happen in the world that are external to us. We observe those things, and that observation is what I'm calling an experience. We can say things about the experience, but those words are not the experience. As to what the experience maps to, I think the simplest answer is that our phenomenal experiences are encoded as structures in our brain, but that's not necessary to understanding the differen…

Ok, what kind of information structure is that experience encoded in? This is where it's really easy to start thinking the brain is some kind of interesting magic rather than encoded information.

I'm not sure I understand your question but I'll try to answer as best I can - also keep in mind that this is simply one view. The structure of the brain encodes information based on experience in the same way that the force of gravity encodes information on two rocks that collide, or other physical forces encode information into chemical structures, etc.

In the case of the brain the encoding is such that various functions "fall out of" it, like being able to relate experiences, etc.

There's no magic proposed here, this is a physicalist functionalist view.

Nothing about this prevents a computer from being sentient. As I said, none of this even matters. The key premise is that LLMs are trained on language and not experiences. Unless you believe that a description of an experience is identical to the phenomenal experience, then we agree on the key premise. Do you think that they are identical?

Re: Even 'uncensored' models can't say what they want

#149

Post by suburban_strike »

subscribed wrote: post #146
throwawaypath wrote: post #130

Earlier quoted context omitted.

[flagged]

"too effeminate"? I'm trying to understand what you're meaning by it and what the example of the excluded data would be. Mind you, I didn't go through The Pile yet, need ot make some disk space first, so I can't even say if your "accusation" is correct.

Reasoning by emotional manipulation rather than facts, data, or logic.
Post reply on HN