I’m with the people pushing back on the “confidence scores” framing, but I think the deeper issue is that we’re still stuck in the wrong mental model.
It’s tempting to think of a language model as a shallow search engine that happens to output text, but that metaphor doesn’t actually match what’s happening under the hood. A model doesn’t “know” facts or measure uncertainty in a Bayesian sense. All it really does is traverse a high‑dimensional statistical manifold of language usage, trying to produce the most plausible continuation.
That’s why a confidence number that looks sensible can still be as made up as the underlying output, because both are just sequences of tokens tied to trained patterns, not anchored truth values. If you want truth, you want something that couples probability distributions to real world evidence sources and flags when it doesn’t have enough grounding to answer, ideally with explicit uncertainty, not hand‑waviness.
People talk about hallucination like it’s a bug that can be patched at the surface level. I think it’s actually a feature of the architecture we’re using: generating plausible continuations by design. You have to change the shape of the model or augment it with tooling that directly references verified knowledge sources before you get reliability that matters.
>A model doesn’t “know” facts or measure uncertainty in a Bayesian sense. All it really does is traverse a high‑dimensional statistical manifold of language usage, trying to produce the most plausible continuation.
And is that that different than what we do under the scenes? Is there a difference between an actual fact vs some false information stored in our brain? Or both have the same representation in some kind of high‑dimensional statistical manifold in our brains, and we also "try to produce the most plausible continuation" using them?
There might be one major difference is at a different level: what we're fed (read, see, hear, etc) we also evaluate before storing. Does LLM training do that, beyond some kind of manually assigned crude "confidence tiers" applied to input material during training (e.g. trust Wikipedia more than Reddit threads)?
I would say it's very different to what we do. Go to a friend and ask them a very niche question. Rather than lie to you, they'll tell you "I don't know the answer to that". Even if a human absorbed every single bit of information a language model has, their brain probably could not store and process it all. Unless they were a liar, they'd tell you they don't know the answer either! So I personally reject the framing that it's just like how a human behaves, because most of the people I know don't lie when they lack information.
>Go to a friend and ask them a very niche question. Rather than lie to you, they'll tell you "I don't know the answer to that"
Don't know about that, bullshitting is a thing. Especially online, where everybody pretends to be an expert on everything, and many even believe it.
But even if so, is that because of some fundamental difference between how a human and an LLM store/encode/retrieve information, or more because it has been instilled into a human through negative reinforcement (other people calling them out, shame of correction, even punishment, etc) not to make things up?
Solid agree. Hallucination for me IS the LLM use case. What I am looking for are ideas that may or may not be true that I have not considered and then I go try to find out which I can use and why.
This technology (which I had a small part in inventing) was not based on intelligently navigating the information space, it’s fundamentally based on forecasting your own thoughts by weighting your pre-linguistic vectors and feeding them back to you. Attention layers in conjunction of roof later allowed that to be grouped in higher order and scan a wider beam space to reward higher complexity answers.
When trained on chatting (a reflection system on your own thoughts) it mostly just uses a false mental model to pretend to be a desperate intelligence.
Thus the term stochastic parrot (which for many us actually pretty useful)
Thanks for your input - great to hear from someone involved that this is the direction of travel.
I remain highly skeptical of this idea that it will replace anyone - the biggest danger I see is people falling for the illusion. That the thing is intrinsically smart when it’s not - it can be highly useful in the hands of disciplined people who know a particular area well and augment their productivity no doubt. Because the way we humans come up with ideas and so on is highly complex. Personally my ideas come out of nowhere and mostly are derived from intuition that can only be expressed in logical statements ex-post.
Is intuition really that different than LLM having little knowledge about something? It's just responding with the most likely sequence of tokens using the most adjacent information to the topic... just like your intuition.
With all due respect I’m not even going to give a proper response to this… intuition that yields great ideas is based on deep understanding. LLM’s exhibit no such thing.
These comparisons are becoming really annoying to read.
I mean... That is exactly how our memory works. So in a sense, the factually incorrect information coming from LLM is as reliable as someone telling you things from memory.
But not really? If you ask me a question about Thai grammar or how to build a jet turbine, I'm going to tell you that I don't have a clue. I have more of a meta-cognitive map of my own manifold of knowledge than an LLM does.
Hallucinations are a feature of reality that LLMs have inherited.
It’s amazing that experts like yourself who have a good grasp of the manifold MoE configuration don’t get that.
LLMs much like humans weight high dimensionality across the entire model then manifold then string together an attentive answer best weighted.
Just like your doctor occasionally giving you wrong advice too quickly so does this sometimes either get confused by lighting up too much of the manifold or having insufficient expertise.
I asked Gemini the other day to research and summarise the pinout configuration for CANbus outputs on a list of hardware products, and to provide references for each. It came back with a table summarising pin outs for each of the eight products, and a URL reference for each.
Of the 8, 3 were wrong, and the references contained no information about pin outs whatsoever.
That kind of hallucination is, to me, entirely different than what a human researcher would ever do. They would say “for these three I couldn’t find pinouts” or perhaps misread a document and mix up pinouts from one model for another.. they wouldn’t make up pinouts and reference a document that had no such information in it.
Of course humans also imagine things, misremember etc, but what the LLMs are doing is something entirely different, is it not?
Humans are also not rewarded for making pronouncements all the time. Experts actually have a reputation to maintain and are likely more reluctant to give opionions that they are not reasonably sure of. LLMs trained on typical written narratives found in books, articles etc can be forgiven to think that they should have an opionion on any and everything. Point being that while you may be able to tune it to behave some other way you may find the new behavior less helpful.
> Hallucinations are a feature of reality that LLMs have inherited.
Huh? Are you arguing that we still live in a pre-scientific era where there’s no way to measure truth?
As a simple example, I asked Google about houseplant biology recently. The answer was very confidently wrong telling me that spider plants have a particular metabolic pathway because it confused them with jade plants and the two are often mentioned together. Humans wouldn’t make this mistake because they’d either know the answer or say that they don’t. LLMs do that constantly because they lack understanding and metacognitive abilities.
>Huh? Are you arguing that we still live in a pre-scientific era where there’s no way to measure truth?
No. A strange way to interpet their statement! Almost as if you ...hallucinated their intend!
They are arguing that humans also hallucinate: "LLMs much like humans" (...) "Just like your doctor occasionally giving you wrong advice too quickly".
As an aside, there was never a "pre-scientific era where there [was] no way to measure truth". Prior to the rise of modern science fields, there have still always been objective ways to judge truth in all kinds of domains.
When you ask humans however there are all kinds of made-up "facts" they will tell you. Which is the point the parent makes (in the context of comparing to LLM), not whether some legal database has wrong cases.
Since your example comes from the legal field, you'll probably very well know that even well intentioned witnesses that don't actively try to lie, can still hallucinate all kinds of bullshit, and even be certain of it. Even for eye witnesses, you can ask 5 people and get several different incompatible descriptions of a scene or an attacker.
That’s deliberate. “Correct” implies anchoring to a truth function the model doesn’t have. “Plausible” is what it’s actually optimising for, and the disconnect between the two is where most of the surprises (and pitfalls) show up.
As someone else put it well: what an LLM does is confabulate stories. Some of them just happen to be true.
Do you have a better word that describes "things that look correct without definitely being so"? I think "plausible" is the perfect word for that. It's not a sleight of hand to use a word that is exactly defined as the intention.
They are, the model has no inherent knowledge about its confidence levels, it just adds plausible-sounding numbers. Obviously they _can_ be plausible, but trusting these is just another level up from trusting the original output.
I read a comment here a few weeks back that LLMs always hallucinate, but we sometimes get lucky when the hallucinations match up with reality. I've been thinking about that a lot lately.
> the model has no inherent knowledge about its confidence levels
Kind of. See e.g. https://openreview.net/forum?id=mbu8EEnp3a, but I think it was established already a year ago that LLMs tend to have identifiable internal confidence signal; the challenge around the time of DeepSeek-R1 release was to, through training, connect that signal to tool use activation, so it does a search if it "feels unsure".
Wow, that's a really interesting paper. That's the kind of thing that makes me feel there's a lot more research to be done "around" LLMs and how they work, and that there's still a fair bit of improvement to be found.
In science, before LLMs, there's this saying: all models are wrong, some are useful. We model, say, gravity as 9.8m/s² on Earth, knowing full well that it doesn't hold true across the universe, and we're able to build things on top of that foundation. Whether that foundation is made of bricks, or is made of sand, for LLMs, is for us to decide.
G, the gravitational constant is (as far as we know) universal. I don't think this is what they meant, but the use of "across the universe" in the parent comment is confusing.
g, the net acceleration from gravity and the Earth's rotation is what is 9.8m/s² at the surface, on average. It varies slightly with location and altitude (less than 1% for anywhere on the surface IIRC), so "it's 9.8 everywhere" is the model that's wrong but good enough a lot of the time.
It doesn't even hold true on Earth! Nevermind other planets being of different sizes making that number change, that equation doesn't account for the atmosphere and air resistance from that. If we drop a feather that isn't crumpled up, it'll float down gently at anything but 9.8m/s². In sports, air resistance of different balls is enough that how fast something drops is also not exactly 9.8m/s², which is why peak athlete skills often don't transfer between sports. So, as a model, when we ignore air resistance it's good enough, a lot of the time, but sometimes it's not a good model because we do need to care about air resistance.
Gravity isn't 9.8m/s/s across the universe. If you're at higher or lower elevations (or outside the Earth's gravitational pull entirely), the acceleration will be different.
Their point was the 9.8 model is good enough for most things on Earth, the model doesn't need to be perfect across the universe to be useful.
