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

OpenAI's new open-source model is basically Phi-5

seangoedecke.com

Post reply on HN
Print view View original

191–200 of 233 posts

Re: OpenAI's new open-source model is basically Phi-5

#191

Post by Gracana »

katzenversteher wrote: post #166
sterlind wrote: post #87

Earlier quoted context omitted.

it's not erotic role-play, but I have a use case of making an AI-powered NetHack clone. specifically, to generate dungeon layouts, dialog for NPCs and to fill in the boatloads of minutae and interactions which NetHack is famous for. you kind of need soul for that, and a lot of background knowledge on mythology/fantasy lore, but also tool use to work the world systems.

I've been experimenting with using various LLMs as a game master for a Lovecraft-inspired role-playing game (not baked into an application, just text-based by prompting). While the LLMs can generate scenarios that fit the theme, they tend to be very generic. I've also noticed that the models are extremely susceptible to suggestion. For example, in one scenario, my investigator was in a bar, and when I commented to an…

Have you used any thinking models? I remember being surprised by QwQ-32B when I tried it. It would think about what I said and how it should respond, reiterate the behaviors I had assigned to it, and respond accordingly. That constant self-reinforcement in the thinking phase seemed to keep it on track.

Re: OpenAI's new open-source model is basically Phi-5

#192

Post by rynn »

Azure AI Foundry says:

How was the model trained? The gpt-oss models were pretrained primarily using synthetic data along with some heavily filtered real code. The models were then post-trained using distillation (RLKD against o3/hailmary) and berry. For more details about the model training process, please refer to the documentation.

Meaning the author was spot on.

Re: OpenAI's new open-source model is basically Phi-5

#195

Post by sdenton4 »

isoprophlex wrote: post #122
zacmps wrote: post #120

Earlier quoted context omitted.

Might be fixable with promoting and/or a much lower temperature?

I've had some success in letting it generate lists of things at high temperature and picking the last item in the list. Ask some model to generate a number uniformly between 0 and 100; you get 47 a lot. Or 27. Something like that. Ask it for a list of numbers, uniformly distributed, and it also often starts with the same number. However later elements in the sequence will converge (imperfectly) to a better approximat…

I have been using one model to generate piles of random-adjective tables, then use real live randomness to select a bunch of traits to feed to the model doing character generation. It's instructed that it can take or leave individual traits.

Re: OpenAI's new open-source model is basically Phi-5

#196

Post by Foobar8568 »

orbital-decay wrote: post #176

I gave it a random sci-fi novel and made it translate a chapter, which is something I do with all models. It refused to discuss minors in sexualized contexts. I was like W.T.F.?! and started bisecting the book, trying to find the piece that triggers this. Turns out there was some absolutely innocent, two sentence long romantic remark involving two secondary 17 years old characters in an unrelated place. Another issue…

The 20B refused to acknowledge that he gave me wrong informations.

Usually models just apologize after I insist 2 or 3 times.

So it was the shortest LLM I tried, I honestly can't trust such models for anything.

Re: OpenAI's new open-source model is basically Phi-5

#197

Post by Havoc »

wkat4242 wrote: post #188
Havoc wrote: post #171

Earlier quoted context omitted.

Pretty sure I saw an uncensored version yesterday on localllama

huh interesting thanks! I'll have a look.

Think they call it abliterated not uncensored but same thing from what I understand

Re: OpenAI's new open-source model is basically Phi-5

#198

Post by bobsmooth »

spacecadet wrote: post #178
orbital-decay wrote: post #176

I gave it a random sci-fi novel and made it translate a chapter, which is something I do with all models. It refused to discuss minors in sexualized contexts. I was like W.T.F.?! and started bisecting the book, trying to find the piece that triggers this. Turns out there was some absolutely innocent, two sentence long romantic remark involving two secondary 17 years old characters in an unrelated place. Another issue…

Its a public consumer facing model, not surprised. Go find an unaligned model that will better produce the content you seek...

After running Dolphin Mistral on my own machine I won't trust censored models ever again.

Re: OpenAI's new open-source model is basically Phi-5

#199

Post by jononor »

Mars008 wrote: post #146
KingOfCoders wrote: post #117

Earlier quoted context omitted.

I don't care if it's open source or not. I care if people call something open source which it isn't. Do you care about binary blobs in the kernel? No. Are binary blobs in the kernel open source? No. But it is tedious to go through the same discussion every 10 years, with a relentless industry that wants to dupe people. If there wasn't a benefit in for them, they would not call it open source.

For some reason people think of models as software and open source should have similar meaning. There are fundamental differences: 1) models aren't reproducible given everything, data, hardware, methodology. 2) they aren't even verifiable. i.e. given model and dataset it's impossible to say if model was trained on that data. 3) except for toys models are trained on copyrighted data. Some of it is private, like users'…

A decent training pipeline will be able to reproduce models with equivalent aggregate performance (as measured by the evaluation metrics). And a high degree of similarity in behavior on specific inputs - but not identical. It will not be the same exact weights. But that is not a critical bar to reach. And may software builds also fails that bar.

Re: OpenAI's new open-source model is basically Phi-5

#200

Post by antonvs »

slt2021 wrote: post #58
kristopolous wrote: post #25

Earlier quoted context omitted.

Porn is always the frontier. It's a well-understood self-contained use-case without many externalities and simple business models. What more, with porn, the medium is the product probably more than the content. Having it on home-media in the 80s was the selling point. Getting it over the 1-900 phone lines or accessing it over the internet ... these were arguably the actual product. It might have been a driver of earl…

even if it is victim free, it can affect mental health in a way that a consumer will be more compelled to do a criminal act and create a real victim. let's say you publish a Steam game how to be a school shooter and shoot kids, wouldn't that lead to real school shootings ? who can definitely say that computer generated content about criminal behavior, won't lead to real crime with real victims? https://en.wikipedia.o…

Found Tipper Gore’s HN account.
Post reply on HN