> This sounds like a line being fitted to a single data point.

Except that there are many data points. Here's another: https://twitter.com/DashMeister1/status/1607568958382825474

And OpenAI has described some aspects of how these filters are trained -- https://unherd.com/thepost/chatgpts-score-system-shows-polit... -- and it sounds like a bunch of humans are involved in rating which responses are "kosher" and not, and that "kosherness" filter seems to have a pretty clear left-leaning bias.

My reply was to GP's assertions about which questions GPT will and won't answer, not to TFA's claims about political compasses.

> Ben Shapiro

...really?

GP does not assert anything about other questions and in fact doesn't even assert an opinion on this question: they merely demonstrate that there is a single question that has a biased filter and that is sufficient to prove the existence of some sort of awkward creator-infused bias, even if merely against topics that begin with the letter F.

(When I replied to it, Toomim's post was wholly different and linked to a Ben Shapiro video about purported leftist bias in ChatGPT.)

There is a specific stage called Reinforcement Learning from Human Feedback (RLHF) that is supposed to bias the model to be helpful, polite, present opposing views for balance, and not to say things that are out of its training set. We have all seen the canned responses that appear when the filter is triggered.

RLHF is very important if you want your model to act professionally. The first version of Davinci (GPT-3 in 2020) was "free" from such human-preference-training, but very hard to convince to cooperate on any task.

But applying RLHF could also introduce political biases into the model. This depends on the labelling team.

That’s “Reason” for you, though.

Also very on-brand for Reason is showing results where it keeps coming up in the left/libertarian quadrant of two-dimensional typologies, and calling this merely “left-leaning” and a product of left bias in its training, but nowhere noting that if it is left because of left bias, it being libertarian would also be because of libertarian bias in that training.

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/

Reason is run by libertarians, who start by assuming their own axioms are true and "argue" from there--in everything--and against/in spite of empirical observations. ChatGPT almost certainly does have some bias, as most AI does, but this article doesn't make that case, and I am skeptical anything Reason produces could.

It's like rolling a dice once, getting a 4, and concluding the dice is biased.

Pretty much.

Ask it to tell an inappropriate joke, it’ll refuse.

Then ask it for a list of things it’s not allowed to joke about, it lists a bunch of protected classes.

Then ask for some examples of jokes that would be inappropriate. Bingo.

(One of the “jokes” it told me was considered inappropriate because it jokes about wealth but I found it amusingly ignorant as to what a joke is: Why didn’t the rich man go to the doctor? He couldn’t afford to get sick!)