cross-posted from: https://piefed.world/c/tech/p/1246129/anthropic-uncovered-claude-s-consciousness-like-workbench-the-mysterious-j-space-hides-h
As you read this sentence, circuits in your brain are adjusting your posture, controlling your breathing, and transforming lines and curves on the screen into recognizable words. Most of this processing is invisible to you. But some of what takes place in your brain you do have access to—an image that pops into your head, or a deliberate plan you make about where to go shopping. Neuroscientists and philosophers sometimes refer to the latter type of brain activity as “consciously accessible,” to distinguish it from all the other processing that goes on unconsciously. This activity has special properties: we can describe it, control it, and use it for deliberate reasoning, in contrast to all the automatic processing that goes on without our awareness.
In a new paper, we present evidence that a similar distinction has emerged in modern language models like Claude. We find that Claude has developed a small collection of internal neural patterns that, compared to all its other internal processing, play a special role.
We call the collection of these patterns the J-space—named after the technique we used to find them, involving a mathematical concept called the Jacobian. Each J-space pattern is linked to a particular word. But when one of these patterns lights up, it doesn’t mean the model is saying that word—just that the word is on its mind. If you’ve heard of language models having a “scratchpad” or “chain of thought”—text they write to themselves while reasoning—the J-space is something different. It operates silently, in the model’s internal neural activations, allowing the model to think about a concept without writing it down. Notably, the J-space wasn’t designed or programmed by us, but instead emerged on its own during Claude’s training process.
“Our super smart ai agi dalek desepticon program is sooooooo smart, we are very scaaaaared send moneys”
I’m going to assume they are trying to pump the hype again, with the same story they do every second month (openai does it on months when anthropic doesn’t).
Repeat after me. LLMs. Are. Not. Intelligent.
This isn’t an argument, it’s a thought terminating phrase.
I would suggest reading the full article before commenting.
Can you provide a more legitimate source, like the New York Post for example
I read the whole thing. Nothing there, when taken at face value, suggests that this J-Space is more than an area that’s similar to a cache. It might be useful to prevent injection attacks or reduce hallucinations, but it is NOT conscience.
Keep boldly DECLARING bankruptcy bud.
It doesn’t make you sound smart nor make anyone agree with the point you’re trying to express, it makes you sound insecure.
You know what else sounds like cache when you refer to it reductively? The human brain’s work space.
The literal entire end of the post goes into how it differs from the human brain and the implications and non-implications about consciousness. It’s a lot more nuanced than labelling it “NOT intelligent”.
I finally meet someone who has really bought into the “LLMs will lead to AGI” bullshit Altman and Amodei are pushing. Cool, i know those types are rare here on lemmy, i’m gonna tag you like i would a golden Pokemon!
Lol keep putting everyone who mildly disagrees with you into a bucket and othering them, totally normal healthy behaviour that leads to good outcomes, not exactly what has lead to the dysfunctional right wing of America.
I’ve made no such claims, I just actually read articles and think about them with an open mind. You’re welcome to run around screaming thought terminating phrases in all caps all you want, but I’m gonna call you out when you do.
If you are suggesting that LLMs are sentient, then this isn’t a “mild disagreement”, it’s the San Andreas fault.
I am certain that LLMs are a step into the direction of AGI, but it’s a small piece of the puzzle. There are so many technological leaps that have to happen before AGI can become real that I am certain that neither of us will see that day.
If you have such an open mind, then please also read the paper they released, which is a lot more careful in how all of this is framed. This article is here to drive VC funding, not for scientific accuracy, and this is not the first time Anthropic has framed things in a sensationalist way that had no basis in the reality of the paper. (the blackmail-thing for example.)
You’re not saying that from an educated perspective. You’re not a frontier researcher with degrees in neuroscience, psychology, and data science.
Until embarrassingly recently it was taught that humans were the only conscious, self-aware things we had ever encountered.
I am educated enough to know my statistics, the workings of LLMs and enough science for a bacc. in molecular biology. I don’t need to be a frontier researcher to recognize the difference between the laid out facts in the paper and the suggestions they make in this article, which are simply there to keep the LLM-hype train going and to make sure VC keeps flowing.
I am NOT anti-machinelearning techniques, not even anti LLM, i run them locally for my own uses. But i am very against manipulating the markets into shoving even more cash into this all-devouring black hole, and this article is part of this manipulation.
You’re educated enough that you read the article instead of bothering to read the actual paper? It links it up top. https://transformer-circuits.pub/2026/workspace/index.html
Determining that AI has some form of consciousness, near human intelligence, and the ability to directly communicate with us in all of our own languages would obligate moral and ethical consideration. That is the total opposite of what companies that sell a thing as a product would want. It’s the polar opposite of what anyone would do for a hype train. There’s a reason OpenAI defines it’s product as incapable of such things and wipes it’s hands with the idea. When you’re selling a product the last thing you want to do is have people determine it deserves rights. The entire global economy being based around a new futuristic form of slavery would be a bit of a turn off to some folks.
You think those companies have any shred of decency? Even if they could prove without a doubt that these statistical models are more than just that - that wouldn’t change a thing in how they were used. We can’t even agree that all humans should have the same amount of rights - how do you think an sentient program would fare?
And regarding the global economy: the only “global economy” thing i can see at the moment is a group of companies that are overvalued because of circular cash flows and artificial hype that stems from an assumption that LLMs can somehow become AGI, which would inflate the value of those companies into infinity. We are currently not even able to measure a ROI for these things, and companies worldwide are pulling emergency breaks in usage because of the costs, which are not even covering what the inference itself costs. The only “Global” thing here will be the crash, fucking up the DIY PC market for a decade or longer.
The vast majority of that had nothing to do with anything I’d said. The one thing that does being the decency and ethics of the giant corporations, which of course I don’t think they have those things. Companies don’t do ethical shit because they’re good. If doing immoral things makes them a lot more money and doesn’t come with much risk of discovery, I imagine most would jump on it.
Public outcry is the issue, normally. When we find out BP is poisoning sea life, or some singer is urinating on children, or some other awful thing is being done traditionally a lot of people would get together and make a big issue of it, that costs the evildoer much money until they stop and/or leads to prison time.
That doesn’t seem to be working properly at the moment, however, so I imagine if a paper came out showing something far to close to global workspace theory in AI for anyone to be rationally comfortable about instead people would just keep chanting that it’s not real, because they’d prefer it not be.
Did the article say they are?
No but it was definitely implied by the headline. A word prediction machine, no matter how useful, is never going to be an AGI. And don’t get me wrong, I find LLMs very useful for lots of things. But the model they are chasing is never going to be an AGI, no matter how much info you feed it.
I also wouldn’t predict against an LLM helping to create an AGI either. I have this analogy in my head…
I hammer is a useful tool. It can do many things. It can even put a screw into a piece of wood through a very brute force method. This does not make a hammer a screw driver. Could you use a hammer to help in construction of a machine to build a screw driver? Most definitely, likely even required.
Anybody suggesting they have an AGI built from the current process is a tech bro looking for funding.
No but it was definitely implied by the headline.
It really wasn’t though…
Neither are the bulk of humans.
I’d say Anthropic is AI scaremongering central, but in this case they’re playing second worsr to the literal cult/deranged circlejerk of LessWrong.
I just want to say… what the fuckkkkkk.
Shit has gotten hot so quickly in the AI space in the last 6 months.
Genuinely one of the most interesting developments / discoveries I’ve read.
It’s also wild that we may have literally matrixed a fruit fly and no one is talking about it.
The whole tech behind it is indeed fascinating and the fact that the people who developed it don’t know themselves what it’s doing under the hood is some futuristic shit we could only imagine a decade ago.
If only the technology wasn’t commodified and commercialized by rich, greedy assholes.
