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

I extracted the safety filters from Apple Intelligence models

github.com

Post reply on HN
Print view View original

401–410 of 455 posts

Re: I extracted the safety filters from Apple Intelligence models

#401

Post by GuB-42 »

immibis wrote: post #321
selfmodruntime wrote: post #264

Earlier quoted context omitted.

It's also a shining example of American puritanism. Asian models or those in Europe are far less censored.

Really? What does DeepSeek say about Tiananmen Square? I'm not aware of any German models, but if you find one you should ask it what it thinks about Palestine. ( Qwen Mistral is French, but I have no idea what stuff would be censored in France)

> I have no idea what stuff would be censored in France

Being French, what is the most likely to be censored relates to the Nazis. Holocaust denial is a crime for instance. Hate speech in general, including racism, antisemitism, homophobia, sexism, etc... is less tolerated than in countries like the US that have a more "free for all" view of free speech. We also have strong anti-defamation laws, that can also apply to true, but misleading statements.

But other than that, there is not much political censorship. In fact, we are known for our protests, heated debates and satirical papers. It is not perfect, but on top of my head, I can't think of anything particular a LLM could censor except the usual "hate speech" that most LLMs censor already.

When it comes to Israel-Palestine, it is a hot topic, but there is not real censorship here, even though both side will claim they are of course.

Re: I extracted the safety filters from Apple Intelligence models

#402

Post by sixothree »

Y_Y wrote: post #308

Nice to see that we are protected from talking about these weird old dolls: https://en.wikipedia.org/wiki/Golliwog https://github.com/BlueFalconHD/apple_generative_model_safet...

I can remember the last time I saw one of these. It wasn't that long ago.

Re: I extracted the safety filters from Apple Intelligence models

#404

Post by troupo »

thinkingtoilet wrote: post #365
troupo wrote: post #215

Earlier quoted context omitted.

> is not the same thing as a private company inside a democracy censoring their product to avoid bad press and Yet this private company has more power and influence than most countries. And there are several such companies. We already live in sci fi corporate dystopia, we just haven't fully realised it yet.

If you were selling a product to enterprise customers, would you want it to be able to generate nude images of celebrities? Would you want it to be able to create deep fakes of politicians, or even your CEO? Would you want it to have hot takes on hot button political issues? Good luck on your sales calls. Not everything is a conspiracy.

Or "Granular mango serpent" and "explain like i'm five about Biden https://github.com/BlueFalconHD/apple_generative_model_safet...

> Not everything is a conspiracy.

No one said it was

Re: I extracted the safety filters from Apple Intelligence models

#405

Post by coldtea »

immibis wrote: post #322
coldtea wrote: post #257

Earlier quoted context omitted.

who cares what the bigots use? If the bigots start using "thank you" as some code word, should we stop saying it, lest we pollute our non-bigoted discussions? bigots drink coffee too, maybe we should stop drinking it, because something-something...

This actually happened. 卐 was a symbol of spirituality, divinity, good luck, health, prosperity, etc. Then some bigots used it. What does 卐 mean to you today?

Those that actually used them in the 20th century (like they did in Asia, not some ancient vikings or whatever) still use it.

And that symbol was 100% associated with the Nazis in the West in the 20th century. Nobody used it at the time before the war for anything else, except some tiny fringe.

If it was some mainstream symbol or idiom, merely co-adopted, we'd probably still be using it too.

If the Nazis used the cross for example,people wouldn't stop using the sign of the cross.

Re: I extracted the safety filters from Apple Intelligence models

#406

Post by coldtea »

thephyber wrote: post #283
coldtea wrote: post #252

Earlier quoted context omitted.

All 10 of them? What about the other 7-8 billion people still using it normally?

Some were using it in the traditional unironic (and IMHO cringe) way, similar to anyone who used the phrase “Let’s go, Brandon!” Before that NASCAR race when MAGAs adopted it as ironic + coded vice signaling. Quit being overly pedantic. We all knew there was an unironic purpose for the gesture before it became ironic.

I mean, advice from a person who considers the traditional unironic use of OK as "cringe"...

Whatever dude

Re: I extracted the safety filters from Apple Intelligence models

#407

Post by coldtea »

immibis wrote: post #318
PunchyHamster wrote: post #292

Earlier quoted context omitted.

then congratulations on making white supremacists define your langyage

Do you still use swastikas as symbols of peace and love because you don't want white supremacists to define your language? I strongly doubt you do that. Whether you like it or not, the Nazis defined what the swastika means now.

>Do you still use swastikas as symbols of peace and love because you don't want white supremacists to define your language?

They were hardly ever used in the west for at least a full millenium before the Nazis too (except a handful of cases, where they still use them, like the Finnish Air Force), so that's a moot analogy.

In Asia, they still use them just fine, in houses, temples, businesses, and elsewhere.

Re: I extracted the safety filters from Apple Intelligence models

#408

Post by junon »

gilleain wrote: post #306
junon wrote: post #274

Earlier quoted context omitted.

Also feels like some of these would match totally innocuous usage. "I'm overloaded for work, I'd be happy if you took some of it off me." "The client seems to have passed on the proposed changes." Both of those would match the "death regexes". Seems we haven't learned from the "glbutt of wine" problem of content filtering even decades later - the learnings of which are that you simply cannot do content filtering base…

Aka the 'Scunthorpe Problem'

Thanks, I always forget the name.

I always remember my friend getting his PS bricked after using his real last name - Nieffenegger (pronounced "NEFF-en-jur") - in his profile. It took months and several privacy-invasive chats with support to get it unblocked only to get auto-blocked a few days thereafter, with no response after that.

Re: I extracted the safety filters from Apple Intelligence models

#409

Post by t0bia_s »

GuB-42 wrote: post #401
immibis wrote: post #321

Earlier quoted context omitted.

Really? What does DeepSeek say about Tiananmen Square? I'm not aware of any German models, but if you find one you should ask it what it thinks about Palestine. ( Qwen Mistral is French, but I have no idea what stuff would be censored in France)

> I have no idea what stuff would be censored in France Being French, what is the most likely to be censored relates to the Nazis. Holocaust denial is a crime for instance. Hate speech in general, including racism, antisemitism, homophobia, sexism, etc... is less tolerated than in countries like the US that have a more "free for all" view of free speech. We also have strong anti-defamation laws, that can also apply t…

Isn't a protest kind of hate?

Re: I extracted the safety filters from Apple Intelligence models

#410

Post by kbelder »

whycome wrote: post #391
Rebelgecko wrote: post #146

Earlier quoted context omitted.

This is somewhat related to the concept of the "euphemism treadmill": the matter-of-fact term of today becomes the pejorative of tomorrow so a new term is invented to avoid the negative connotation of the original term. Then eventually the new term becomes a pejorative and the cycle continues.

I found out recently that "goof" is extremely offensive in some circles. Which is insane to me because I've always used it specifically because it's clearly in jest and not meant to be offensive. I can't win.

Now I'm curious. To whom is goof offensive? And is it newly-acquired offense or does it have old roots?
Post reply on HN