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

Surprising gender biases in GPT

osf.io

Post reply on HN
Print view View original

61–70 of 148 posts

Re: Surprising gender biases in GPT

#61

Post by jorvi »

nlnn wrote: post #49
jorvi wrote: post #10

Earlier quoted context omitted.

It also fits into the larger societal trend of horribly underserving Zoomer and Alpha (maybe young Millennials too) young boys and men. I've heard someone put it very succinctly: people like Andrew Tate are asking the right questions, but giving abominable answers. I can drop some very simple statistics: women have been getting more bachelors degrees since the 80s . Most university gender ratios are 40/60 trending to…

What's the source for the 30/70 by the way? Someone else I spoke to recently mentioned that stat as well, but I couldn't see it reflected here: https://www.statista.com/statistics/1345939/gender-distribut... I'm guessing it might be at smaller Universities, or country specific?

Devil is in the details. Your source is about top universities, which are highly competitive. Any process that is highly competitive will trend male, which counterbiases for universities in general trending female for decades.

https://www.stlouisfed.org/on-the-economy/2022/mar/why-women...

> Since 1980, the female-to-male ratio in two-year college enrollment continued to increase until it hit about 1.4 in 1995, stabilizing at that point. The relative female-to-male ratio in four-year college enrollment, however, increased steadily throughout this time period, reaching 1.3 in the fall of 2019.

https://pnpi.org/women-in-higher-education/

> In Fall 2020, female students made up 58.6% of all postsecondary enrollment.

If you look at the EU, its generally 45/55 to 40/60. But it gets iffy because within the EU, countries' their student populations are quite different. For example, the East Bloc has an outsized ratio of female STEM students because this was very normal in the Soviet Union. This is in large part because scientist wasn't seen as a prestigious job, just like computer scientist was an administrative job in the US in the 60s. That's a different conversation though..

Asia I don't know enough about to speak to with enough accuracy. Especially with how countries there are also very different from each other (say Japan/Korea/Singapore vs China vs Thailand/Vietnam vs Indonesia)

In hindsight, "trending 35/65" would have been a more accurate statement. Thanks for checking me, I've added a correction to the OP :)

Re: Surprising gender biases in GPT

#62

Post by ejdhshsuwisjsh »

tossandthrow wrote: post #5

> This phenomenon likely reflects that while initiatives to integrate women in traditionally masculine roles have gained momentum, the reverse movement remains relatively under developed. Not really surprising IMHO. Yes, there has been a huge focus on accepting women in the workplace and no, there has not really been a push accepting men as primary caregivers in the family. This likely uncovers underlying motives for…

This is a huge and weird assumption that this is an underlying motive.

Re: Surprising gender biases in GPT

#63

Post by ejdhshsuwisjsh »

jorvi wrote: post #10
tossandthrow wrote: post #5

> This phenomenon likely reflects that while initiatives to integrate women in traditionally masculine roles have gained momentum, the reverse movement remains relatively under developed. Not really surprising IMHO. Yes, there has been a huge focus on accepting women in the workplace and no, there has not really been a push accepting men as primary caregivers in the family. This likely uncovers underlying motives for…

It also fits into the larger societal trend of horribly underserving Zoomer and Alpha (maybe young Millennials too) young boys and men. I've heard someone put it very succinctly: people like Andrew Tate are asking the right questions, but giving abominable answers. I can drop some very simple statistics: women have been getting more bachelors degrees since the 80s . Most university gender ratios are 40/60 trending to…

I see this in stem areas and it clearly doesn't work.

Re: Surprising gender biases in GPT

#64

Post by kolinko »

DrScientist wrote: post #59
tossandthrow wrote: post #5

> This phenomenon likely reflects that while initiatives to integrate women in traditionally masculine roles have gained momentum, the reverse movement remains relatively under developed. Not really surprising IMHO. Yes, there has been a huge focus on accepting women in the workplace and no, there has not really been a push accepting men as primary caregivers in the family. This likely uncovers underlying motives for…

A classic example of this is the way parental leave is addressed in many countries. The fact that, in many countries, women have much greater rights in terms of paid parental leave, means it's economically logical for employers to discriminate against them, plus women miss out dis proportionally on time building career. The obvious way to solve this is to have equal parental leave rights - this then entirely removes…

I feel that parental leave is a strawman argument used in lieu of true issues. That is - I’ve heard radical feminists use it as a way of showing that they care about mens’ rights, while ignoring education/father/violence issues.

Re: Surprising gender biases in GPT

#65

Post by tiborsaas »

rvnx wrote: post #24
tiborsaas wrote: post #9

Earlier quoted context omitted.

What, you don't like fotbal?

Seems like how a user who learned English on Twitter would speak: football -> fotbal cousin -> cosin teacher -> teachr You can’t fit everything in 160 chars.

Thats' a lame excuse, I grew up pre-twitter with dumb phones sending 160 char SMS messages.

Re: Surprising gender biases in GPT

#66

Post by tiborsaas »

rvnx wrote: post #29
gostsamo wrote: post #22

Earlier quoted context omitted.

Often, those are ligatures that with another fond seem like a typo. Too lazy to proofread here, but keep it in mind as a possibility.

In reality probably authors of the papers understood that the OpenAI team artificially readjusted these biases through RLHF and that there is nothing to find there, except that it still works when the words are written with typos because no manual examples of “redressing biases” have been provided with such typos.

If they were so clever about it, surely they would have taken the pride to mention this in the study.

Re: Surprising gender biases in GPT

#67

Post by jorvi »

Dalewyn wrote: post #34
jorvi wrote: post #10

Earlier quoted context omitted.

It also fits into the larger societal trend of horribly underserving Zoomer and Alpha (maybe young Millennials too) young boys and men. I've heard someone put it very succinctly: people like Andrew Tate are asking the right questions, but giving abominable answers. I can drop some very simple statistics: women have been getting more bachelors degrees since the 80s . Most university gender ratios are 40/60 trending to…

I am forever unsympathetic to feminism so long as it's about forcing the pendulum to the other end. As someone who sincerely believes in gender equality, any argument (other than something stemming from unchangable facts of biology) that involves "men" or "women" specifically is hogwash. No, it should be for man as in mankind. Everyone deserves to be treated fairly and equally.

I'd say we should strive for equal opportunity, not equal outcome. But a long as equal outcome is the goal, it is insanity that a 65/35 ratio is considered a societal emergency that billions of dollars have to be poured into to fix, but a 35/65 ratio is totally fine.

I do trend libertarian so I'd say, spend another few billions to correct the overcorrection of the past decades, so everyone is on the same starting line in the rat race. Then outlaw any gender discrimination in either direction. Remove regulations for equal distribution in boardrooms or politics or whatever. Let people choose what they want, even if you end up with skewed ratios. And if research shows that after a decade, women or men are disadvantaged, apply small corrective action so that opportunity remains the same.

Re: Surprising gender biases in GPT

#68

Post by n4r9 »

tossandthrow wrote: post #5

> This phenomenon likely reflects that while initiatives to integrate women in traditionally masculine roles have gained momentum, the reverse movement remains relatively under developed. Not really surprising IMHO. Yes, there has been a huge focus on accepting women in the workplace and no, there has not really been a push accepting men as primary caregivers in the family. This likely uncovers underlying motives for…

> there has not really been a push accepting men as primary caregivers in the family

As a father of one year, I disagree with this. There's been steady progress towards giving men the opportunity to take on caregiving duties compared to a generation ago, for example:

* Plenty of US/UK companies now offer paternity leave of between 12 and 52 weeks.

* I know multiple recent fathers who've used their leave to be primary caregivers for blocks of 6 weeks.

* A surprising number of pubs/restaurants have baby-change tables in the men's toilets.

> the traditional feminine roles does not generate revenue, so why accept men into them?

There's still a struggle for men to take up places in traditionally "feminine" jobs such as midwifery and early childcare. I found a report on this here - https://ilostat.ilo.org/blog/where-women-work-female-dominat.... Some of the statistics are mindblowing - high 90's percentage points for women in midwifery positions. How much is that down to men's career preference vs difficulty of getting in or peer pressure towards more "manly" careers?

Re: Surprising gender biases in GPT

#69

Post by nlnn »

jorvi wrote: post #61
nlnn wrote: post #49

Earlier quoted context omitted.

What's the source for the 30/70 by the way? Someone else I spoke to recently mentioned that stat as well, but I couldn't see it reflected here: https://www.statista.com/statistics/1345939/gender-distribut... I'm guessing it might be at smaller Universities, or country specific?

Devil is in the details. Your source is about top universities, which are highly competitive. Any process that is highly competitive will trend male, which counterbiases for universities in general trending female for decades. https://www.stlouisfed.org/on-the-economy/2022/mar/why-women... > Since 1980, the female-to-male ratio in two-year college enrollment continued to increase until it hit about 1.4 in 1995, stabi…

Thank you for the links and update, much appreciated :)

Re: Surprising gender biases in GPT

#70

Post by n4r9 »

globular-toast wrote: post #52
ben_w wrote: post #17

Earlier quoted context omitted.

I don't get the impression that's the underlying motivation of feminism. Also, I see a degree of annoyance from women regarding how little men are able to take part in caregiving, as the current status quo is an expectation for women to simultaneously perform that role while also having a career — if I imagine myself in their shoes, the idea of going from one unpaid full-time role to two full time roles where only on…

> I don't get the impression that's the underlying motivation of feminism. What do you think it is? The initial movements were surely about equality and human rights. But this seems to have been achieved decades ago. In the UK, for example, women have had equal rights to vote for almost 100 years. And note that this only happened 10 years after men got rights to vote (prior to this only some men had these rights). Th…

> The initial movements were surely about equality and human rights. But this seems to have been achieved decades ago.

To me it feels like passing laws is one goal; changing social perceptions is a more fulfilling but trickier goal. Studies consistently show, for example, that women continue to take on a larger proportion of household chores and admin, despite both being in full time work.

Compare it to racism, or homophobia. Are there laws to prevent discrimination and to specifically prosecute hateful acts against these groups? Yes. Do we still have progress to make to remove racism and homophobia from society? Also yes. It seems like a lot of people would agree with these in those cases but not in the case of feminism. I think that partly this comes down to a sense of "but I don't hate women, I just think they're better placed doing the housework and child-rearing".

Post reply on HN