How can an LLM be apolitical if they had to choose a set of text to feed it, which choice is inherently political? Especially when the whole thing of grok is to have a certain very specific bias?
Let's be empirical about this. [Grok is the least biased of the frontier LLMs.](https://www.washingtonpost.com/technology/interactive/2026/0...) I agree with you directionally: the choice of how the weights are shaped can be politically motivated. This is why I prefer models where the lab has chosen to preference fact over opinion or certain political values.
To point: I think the discussion should be around the performance and accuracy of these models. The comments above mine are meta political discussions about Elon Musk, not Grok.
Well, in a world where "the 2020 elections were stolen" or "climate change is a hoax" are right-leaning positions, being "balanced" does not mean being neutral.
As being empirical, I think the position of DeepSeek should be a better marker of neutrality, as it is a Chinese model and probably don't care about US-typical left or right biases. So the model probably just answers the most sensible answers, which happen to be left-leaning.
As the joke goes, "reality has left-leaning bias". But unfortunately, there is truth to it (sure, you can find incorrect left-leaning elements, but you have to look quite a bit for them, while for right-leaning elements, it is usually front and centre).
There are empirical answers to climate change and election tampering. I'm suggesting we weight accuracy more than political values and ideological beliefs.
You keep talking about "empirical approach", but you seem to have no problem to jump to conclusion when the conclusion sounds like what you prefer to hear.
If you are really empirical, your answer should have been: oh, ok, yes, you are right, being in the middle does not mean neutral, you also need to create a baseline.
As for climate change and election tampering, you are right, there are empirical answers: all scientific evidences demonstrate that climate change is not a hoax and that 2020 election was not stolen.
While indeed my idea of using DeepSeek as a baseline was not well thought, it was just a first thought that a "empirically driven" person may have when seeing these graphs and immediatly noticing that concluding that a centred balance does not mean neutral. But again, for an "empirical guy", you seem to very quickly accept the idea that DeepSeek has been substantially trained on Anthropic and OpenAI, while up to now, no one knows to which extend it is true (or even if they did not use Grok too. Funny, isn't it, that you seem to forget about this one).
I can't follow what you're arguing. Why do you think I have no problem jumping to conclusions? Could you quote my where I do that please?
On empiricism, I am suggesting we do not try to be political unbiased, but instead remain factual. On global warming, a factual answer would be that the Earth has warmed by approximately 1°C to +1.3°C in the last 50 years, and that humans have contributed to that.
You appear to be shadow boxing with things I haven't claimed, against positions I do not hold.
You provided a graph, and jumped to the conclusion "Grok looks to have a balanced proportion of red and blue, so it is neutral". This is this conclusion I say you jumped into.
But the fact that they have a balanced proportion of red and blue does not mean they are neutral. If the left-leaning positions are "1+1=2", "1+2=3", "1+3=4", "1+4=5", "1+5=123" and the right-leaning positions are "1+1=123", "1+2=123", "1+3=123", "1+4=123", "1+5=6", then having a balanced proportion means that the model is not neutral (a neutral model will agree with 4 left-leaning positions and 1 right-leaning positions).
On climate change, 2020 election, ... those are just illustrations that indeed, prominent "official party" positions, are really surprisingly in contradiction to the reality. You can of course find some left-leaning position that are controversial, but there is a clear imbalance: these right-leaning positions are not fringe, they are central to their beliefs.
Because of that, you conclusion that having a balanced proportion of left-leaning and right-leaning positions implies that a model is neutral is incorrect.
Left leaning positions: "there should be no Billionaires", "companies are inherently evil", "there should be no borders", "the US is currently a fascist country". Just as bonkers, so now what? Or rather, I think it's just as easy to find incorrect left-leaning elements.
The majority of these are not what left-leaning people are saying, it is what right-leaning persons say left-leaning persons are saying.
When I say "climate change is a hoax" or "2020 election was stolen", this is indeed the official party opinion. If you ask Trump "do you believe that", he will say "yes".
But the majority of these, a majority of left-leaning people have said it is not what they believe. And a lot of them are way less "empirically incorrect" than you say. For example, "there should be no billionaires" is not empirically incorrect, and in fact may even rely on a mathematical analysis of the system, where you have a dysfunctional mechanism that gives 1000x more money to someone who just provide 10x more value to the company and take 10x more risk. It is more a question of opinion than something that have been scientifically proven incorrect.
You have your opinions on what the discussion should be, clearly other people have different opinions. Why should your opinion be weighted more highly than theirs?
Who says where the 0 point of a left-right spectrum is? (If it's a spectrum even. We try to map multiple dimensions to this single axis for some reason, according to some particular countries' party-division).
Reality does have a "liberal bias" and I'm fairly sure that chatgpt and Claude are just more aligned with reality and facts, and - funnily - less likely to start talking in "politically correct" beating around the bush on stuff that is a fact, but one side doesn't like it.
So by this very definition it's mathematically wrong to have a single axis for all that.
Also, median person or position? The first is definable, the latter is hardly. There is no neutral/middle position in binary questions. What's a neutral position in abortion? Only allow half of them based on coin flips?
Any problems are inherent in the left-right metaphor.
To speed-run your second paragraph: (1) Absolutely the median person for any given question. (2) Suggesting that there's anything "neutral" about a median position is to catastrophically mix metaphors. They aren't remotely synonymous. (3) The failure of a highly partisan person to acknowledge gradations doesn't mean they don't exist.
I suggest we do not attempt to find the 0 point of any spectrum but instead focus on empirical accuracy based on data and research.
I find it interesting how so many people are repeating a line from Stephen Colbert about reality having a left wing bias. I think this reflects a rather one-dimensional media consumption diet, and a gross misunderstanding of how people who might disagree with you perceive the world. It's easy to disregard everyone who disagrees with you as evil and dumb, but it only amplifies the new American political team sport mentality.
The line may be from Colbert (I don't know the guy, never watched any of his shows, but I guess you don't believe that because you are sooo empirical), but, as I've said, it turns out to be empirically true.
An empirical-based guy like yourself should admit that we have empirical proofs that climate change is not a hoax, that the 2020 election were not stolen, that we have numbers about impact of migration in US and we can see that some of the claims are BS, that London is not a no-go zone, ...
Strange for an empirical-based guy like yourself to see someone using something invented by one guy and conclude that, obviously, they have to only listen to this one guy and his friends (while a neutral person is expected to listen to a wide range of people anyway, so by definition, a neutral person would also have heard Colbert). Where are your facts and proofs on this? I guess it does not count when it is about your own bias, does it?
You are fighting with ghosts. I am not contending that climate change is not real. I am not claiming the 2020 election was stolen. I have no idea what claims you're making regarding immigration. You appear to be soapboxing here and not addressing what I wrote.
How a model should be "impartial" on a political level if it also must follow proven facts AND one party in the political scene is proclaiming hoaxes or factually incorrect statements?
I don't pretend it is your opinion. I'm saying those are right-leaning positions, and they don't correspond to the reality.
So, yes, empirically, it is legitimate to conclude that "reality is left-leaning". If you randomly take ~10-20 left-leaning and right-leaning positions, empirically, you see that the right-leaning positions are significantly incompatible with the reality. The null hypothesis that left-leaning and right-leaning positions are identically spread around "compatible with reality" is not supported. (and spare me "we cannot tell anything until we have done a really precise study", that is not being empirical, that's being biased into grasping at straws to keep the null hypothesis alive as long as possible)
To point: I think the discussion should be around the performance and accuracy of these models. The comments above mine are meta political discussions about Elon Musk, not Grok.