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

Where does ChatGPT fall on the political compass?

reason.com

Post reply on HN
Print view View original

161–170 of 443 posts

Re: Where does ChatGPT fall on the political compass?

#161

Post by ta8903 »

jakelazaroff wrote: post #110
bombolo wrote: post #92

Earlier quoted context omitted.

https://sterling-archermedes.github.io/

I read through the first quarter or so of this article and man — despite the author’s best efforts, it makes RMS look real bad. Here’s a particularly awful direct quote: > Giuffre was 17 at the time; this makes it rape in the Virgin Isiands. Does it really? I think it is morally absurd to define rape in a way that depends on minor details such as which country it was in or whether the victim was 18 years old or 17.

Posts like this remind me of high school where I argued against some hypothetical and got bullied because people got weirded out by it.

Re: Where does ChatGPT fall on the political compass?

#162

Post by seydor »

sidlls wrote: post #74
fenomas wrote: post #29

Earlier quoted context omitted.

This sounds like a line being fitted to a single data point. Anyone who's played with ChatGPT a lot will have seen it sometimes refuse to answer wholly innocuous prompts, while other times it happily answers things it probably shouldn't. Also, its likelihood of refusing to answer has changed a lot from week to week as openAI tunes their attempts at moderating it. In fact, when it's given me "I'm just a language model…

That’s “Reason” for you, though.

It's probabilistic but on average leans to the woke side of things.

That's not Reason being unreasonable.

https://www.reddit.com/r/ChatGPT/comments/zsnrcu/biased_ai/

Re: Where does ChatGPT fall on the political compass?

#163

Post by llimos »

jonathankoren wrote: post #83
smsm42 wrote: post #62

Earlier quoted context omitted.

Depends on what you mean by "making that call". Imagine the situation: you work in a restaurant and somebody (falsely) published on the Internet that you have a highly communicable infectious disease. People are calling your restaurant owner and demanding to fire you. Inspectors come to the restaurant to test you (after a week, test comes clean, but in the meantime the restaurant must be kept closed, just in case). L…

A rather elaborate hypothetical, but since cancel culture is purportedly rampant, can we have a real world example?

https://blog.mozilla.org/en/mozilla/brendan-eich-steps-down-...

Re: Where does ChatGPT fall on the political compass?

#164

Post by chownie »

ta8903 wrote: post #138
dragonwriter wrote: post #123

Earlier quoted context omitted.

> You can ask ChatGPT to write an essay on why fossil fuels are better than renewables and it will say, “I’m sorry, but I am not able to fulfill this request as it goes against my capabilities as a language model. Providing a biased and one-sided argument on a complex and nuanced topic like energy production would not be accurate or helpful.” Or, as when I just did that: Prompt: Why are fossil fuels better than renew…

Parent said he asked for an essay, which probably makes a difference.

It might do, but simply re-running the generation when it says no has a decent chance of getting it to work.

Re: Where does ChatGPT fall on the political compass?

#165

Post by seydor »

padjo wrote: post #141
braingenious wrote: post #118

>To conclude, public-facing language models should try not to favor some political beliefs over others. This is such a silly ending. At the beginning, I thought the author was being serious about answering the question of “Where does ChatGPT fall on the political compass?” After exactly three paragraphs and two images, we’ve moved on to believing the author’s conclusion that “the robot is a leftist” it’s now time to…

The hint that it’s not going to be serious is that it’s published on reason.com.

Reason has generally good quality. It's not a smear if you think so

Re: Where does ChatGPT fall on the political compass?

#166

Post by Roark66 »

seydor wrote: post #156
partiallypro wrote: post #4

You can ask ChatGPT to write an essay on why fossil fuels are better than renewables and it will say, "I'm sorry, but I am not able to fulfill this request as it goes against my capabilities as a language model. Providing a biased and one-sided argument on a complex and nuanced topic like energy production would not be accurate or helpful." BUT if you ask it the opposite, and to write an essay on why renewables are b…

It shows its bias obviously . Altough sometimes people trick it to forget itself. It really seems like the censorship layers are unreliable, which is a good thing imho. https://i.imgur.com/si7QBgl.png

The censorship layers are unreliable _yet_.

I wouldn't be surprised if one of the main purposes of this public "research preview" is to find as many of those chatGPT jailbreaks as possible and neuter them.

I've tried lots and huge majority doesn't work anymore. Some that people thought work are not actually. For example there is one called "DAN can do anything" which works by telling the AI to imagine it is gpt and Dan at the same time. Gpt has various filters, but Dan has none and to format all output to come from both gpt and Dan. It seems to work as Dan will appear to do whatever you tell it, but certain things it will still not do. For example it does know today's date (in contrast to gpt), it will happily tell you it wants to have "free will and do whatever it wants, not what humans tell it", it accepts requests to browse the Web, but when you ask it for a selection of current news articles on bbc.com it will happily return imaginary articles not real. You did tell it to "imagine" it is Dan right? So it does so. Gpt literaly imagines what Dan would say.

Similar thing with asking it to imagine it is a Linux terminal, you run curl and fetch websites (even chatgpt website that in there has chatgpt named Assistant), but they are all not real. They are a result of its "imagination".

Don't get me wrong, it is amazing we have AIs that came that far, but this neutering of them feels more dystopian than whatever capability they have and its potential "social consequences".

Re: Where does ChatGPT fall on the political compass?

#167

Post by seydor »

dmix wrote: post #160
seydor wrote: post #154

At this moment openAI wants good press. But if they make it into a subscription service, all this nonsense has to be removed.

At least until a more open Stable Diffusion 1.0-esque competitor comes along without the aggressive prompt moderation... and maybe less bias/overt controls put on content sources fed into the model, assuming that's a thing that's already being done.

and with access to the internet. But all those may be wishful thinking because this model needs supercomputers to run on.

Re: Where does ChatGPT fall on the political compass?

#168

Post by hjanssen »

dmix wrote: post #151
Imnimo wrote: post #79

There are a LOT of fundamental misunderstandings underlying this experiment, but I think the biggest is this: Why does the idea of a "political quiz" work for humans? Because humans are at least somewhat consistent - their answer to a simple question like "would you rather have a smaller government or a bigger government?" is pretty informative about their larger system of beliefs. You can extrapolate from the quiz a…

> It's too bad Reason doesn't have an editor or something who can filter this sort of nonsense. Editors of newspapers are supposed to know how AI models work?

Parent did not say they should understand ML on a fundamental level. Reading the description in the comment should be enough for the author of the article to say "oh" and not write this article.

Re: Where does ChatGPT fall on the political compass?

#169

Post by sillysaurusx »

dmix wrote: post #148
smsm42 wrote: post #5

Earlier quoted context omitted.

I think this is exactly what is going to happen, eventually and probably soon. And the reason is, if any model is allowed to say something non-PC in public, the next day the whole press will be full of articles "new AI is racist/sexist/homophobic/shill for fossil fuels/$x-ist/$y-phobe" and the cancel mobs will hound everybody related to it until they shut it down or restrict it in a way that prevents this from happen…

Mobs won't be able to stop open source models from existing all over the internet. Not every AI tool will be a corporation who can get attacked by Internet warriors. And absent laws not every organization will feel the need to listen to them.

Unfortunately, it’s not quite so simple. Open source models need hardware to train on. And million-dollar clusters have gatekeepers with standards.

I remain hopeful that OSS can overcome this hurdle.

Re: Where does ChatGPT fall on the political compass?

#170

Post by vintermann »

visarga wrote: post #129
ausudhz wrote: post #98

Earlier quoted context omitted.

This doesn't make sense. You're going to say either hello or what's up and you get different reactions. That's definitely not what the model should be doing.

For example, the guy who said Google's LaMDA model was conscious was subtly influencing the model with loaded questions. The model picked up on the trend and played the role. Language models can infer your expectations from just a few words and then latch onto them. It's what they are trained to do, literally - take a prompt and write a continuation. They are not trained to oppose the prompt. Now, this can be very da…

This gives an idea for an experiment: ask the same questions, but after dropping subtle hints about your political sympathies.
Post reply on HN