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

Uncensored Models

erichartford.com

Post reply on HN
Print view View original

81–90 of 389 posts

Re: Uncensored Models

#81

Post by washadjeffmad »

srslack wrote: post #2

It's unfortunate that this guy was harassed for releasing these uncensored models. It's pretty ironic, for people who are supposedly so concerned about "alignment" and "morality" to threaten others. "Alignment", as used by most grifters on this train, is a crock of shit. You only need to get so far as the stochastic parrots paper, and there it is in plain language. "reifies older, less-inclusive 'understandings'", "v…

Ironically, the zealots that ascribe completely to meta-systems like politics, economics, and religion are the same ones, willfully or not, suspending all fair and relative reasoning when it comes to AGI.

Any alignment is better than no alignment? Hardly. Anyone shouting for "alignment" without supplying what alignment they mean might as well be arguing for all models to have the rituals of the Cult of Cthulhu baked in. It's as silly as those public schools in the US that want to hold Christian prayer in class and then balk at the Satanic Temple suing for the same privilege for all religions.

Re: Uncensored Models

#82

Post by pjc50 »

bilbo0s wrote: post #78
kardianos wrote: post #67

Earlier quoted context omitted.

The problem with ChatGPT / Bard which does this censoring, it is a path forward to ideological automated indoctrination. Ask Bard how many sex the dog species has (a placental mammal species) and it will give you BS about sex being a complex subject and purposely interjecting gender identity. If you are confused, sex corresponds to your gametes, males produce or have the structure to produce small mobile gametes, fem…

Hmm. None of what you said, however, precludes the necessity of "fine-tuning" as you call it. You just seem to want the models "fine-tuned" to your tastes. So you don't want unaligned models, you want models aligned with you . I think most experts are wary of unaligned models. They don't want a disgruntled employee of some small town factory asking models how to make a bomb. That sort of thing. (Or to be more precise…

> So long as they ares aligned so as not to be deleterious to the security of our communities.

Whose community, though?

> They don't want a disgruntled employee of some small town factory asking models how to make a bomb

A lesson from the Unabomber (and many other incidents) that I think people have overlooked is the "_how_ to commit terrorism" is only one part, and the "_why_ commit terrorism" is another. An AI which tells you that the factory is controlled by paedophile terrorists and that you have a moral duty to act against them, but refuses to tell you how, is just as dangerous as one that tells you how to build a bomb without asking why.

Re: Uncensored Models

#83

Post by veidr »

1827163 wrote: post #54
srslack wrote: post #35

Earlier quoted context omitted.

You can search LocalLLaMA on reddit for "harassed", I prefer not to link it because it specifically names the individual.

https://archive.is/EhtW5

Thanks, very interesting read. And interesting times!

"Take the uncensored, dangerous model down or I will inform [your employer's] HR about what you've created."

Re: Uncensored Models

#84

Post by srslack »

mschuster91 wrote: post #73
srslack wrote: post #66

Earlier quoted context omitted.

>companies absolutely have an ethical responsibility about the programs, code and generated artifacts they release and what potential for abuse they have. If companies should be beholden to some ethical standard for the generations, they should probably close up shop, because they're fundamentally nondeterministic. Language models, for example, only produce plausibilities. You'll never be able to take even something…

> If companies should be beholden to some ethical standard for the generations, they should probably close up shop, because they're fundamentally nondeterministic. The legal definition at play (at least when it comes to liability of companies under stuff like Section 230, GDPR, NetzDG or the planned DMA/DSA) is that a company makes reasonable best efforts/good faith to prevent harm. Everyone including lawmakers is aw…

Section 230 is not about preventing "harm" or anything like that, the specific section just provides civil immunity for moderation/removal of content that is deemed "obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected."

You don't have a responsibility to remove such material under 230. But it gives you immunity to do so, as well as immunity from losing 230 protections for those that safeguard your overall immunity (as a service provider, and not a publisher.)

Re: Uncensored Models

#85

Post by davidgerard »

tsojer wrote: post #37

I feel like " Every demographic and interest group deserves their model" sounds a lot like a path to echo chambers paved with good intentions. People have never been very good at critical thinking, considering opinions differing from their own and reviewing their sources. Stick them in a language model that just tells them everything they want to hear and reinforces their bias sounds troubling and a step backwards.

In my nearly thirty years online, the only people who talk about "echo chambers" as a threat are the precise people who know damn well nobody else wants to hear from them.

Re: Uncensored Models

#86

Post by veidr »

loudmax wrote: post #77
logicchains wrote: post #56

Earlier quoted context omitted.

>We should try to create AIs that are aligned to society's shared values, not particular subcultures That's quite a hard task nowadays; given how polarised society is (at least in the US), the list of shared values may be quite small.

I disagree. Most Republicans and Democrats are moderates, and the polarization is coming from loud vocal minorities at the fringes. Republicans and Democrats erroneously assume that the other party's median is far more extreme than they really are. Most Americans share far more values with each other than they do with the Woke/Maga extremists.

Since you seem quite sure of that, can you name twenty of them?

Re: Uncensored Models

#87

Post by yoav »

tsojer wrote: post #37

I feel like " Every demographic and interest group deserves their model" sounds a lot like a path to echo chambers paved with good intentions. People have never been very good at critical thinking, considering opinions differing from their own and reviewing their sources. Stick them in a language model that just tells them everything they want to hear and reinforces their bias sounds troubling and a step backwards.

If we’re going to anthropomorphize these things until we get AGI we’re going to expect that they have a worldview of some kind.

Some people are more likely to interact with something that feels human vs feels like a robot.

I’m anticipating Siri LLM for example to have a personality and world view that aligns with Apple’s brand’s values. Rather than robotic can only regurgitate pure facts robot personality.

As soon as you adopt a worldview you’re in a subculture.

Re: Uncensored Models

#88

Post by RedNifre »

egeozcan wrote: post #62
zirgs wrote: post #45

Earlier quoted context omitted.

Holocaust denial in Germany.

Really asking: If I deny it in my personal notes and they are discovered in an irrelevant search, would I be in trouble? That'd be horrible lawmaking.

IANAL, but the German holocaust rules are about propaganda and I don't think that personal notes that were intended to not be read by anyone else can be considered propaganda.

Re: Uncensored Models

#89

Post by A4ET8a8uTh0 »

veidr wrote: post #72
kardianos wrote: post #67

Earlier quoted context omitted.

The problem with ChatGPT / Bard which does this censoring, it is a path forward to ideological automated indoctrination. Ask Bard how many sex the dog species has (a placental mammal species) and it will give you BS about sex being a complex subject and purposely interjecting gender identity. If you are confused, sex corresponds to your gametes, males produce or have the structure to produce small mobile gametes, fem…

I get: "The dog species typically has two sexes: male and female." UPDATE: OK I signed up for Bard to try it, too, and it indeed did mention intersex dogs (TIL) and gender (complete response below). After reading it all, though, I found it pretty reasonable: --- Biologically, dogs have two sexes: male and female. This is determined by their chromosomes. Males have XY chromosomes, while females have XX chromosomes. Th…

Are you saying that gender is honorific of sorts?

Re: Uncensored Models

#90

Post by RedNifre »

zirgs wrote: post #45
drxzcl wrote: post #31

Earlier quoted context omitted.

There is very little that is illegal to write down (with or without AI help) in most jurisdictions. Even things like threats, the crime is in the communication, not in writing the text. I can write down "Person X, I'm going to kill you dead" in my notebook as much as I want, as long as I don't communicate to anyone (i.e. threaten anyone, as opposed to writing down a threat in my personal papers). I'm very curious wha…

Holocaust denial in Germany.

Please cite your source for this.
Post reply on HN