[HN Gopher] How outdated information hides in LLM token generati...
       ___________________________________________________________________
        
       How outdated information hides in LLM token generation
       probabilities
        
       Author : anjsimmo
       Score  : 108 points
       Date   : 2025-01-10 08:24 UTC (2 days ago)
        
 (HTM) web link (blog.anj.ai)
 (TXT) w3m dump (blog.anj.ai)
        
       | 0xKelsey wrote:
       | > The scenario that I'm worried about, and that is playing out
       | right now, is that they get good enough that we (or our leaders)
       | become overconfident in their abilities and start integrating
       | them into applications that they just aren't ready for without a
       | proper understanding of their limitations.
       | 
       | Very true.
        
         | aprilthird2021 wrote:
         | This is going to happen like in many engineering industries
         | when a cheaper, more likely to be faulty part has been used in
         | replacement of a more expensive, sturdier part. And people will
         | groan but be unable to stop it, unfortunately
        
           | ihsw wrote:
           | It is happening in insurance right now and it is an
           | unmitigated disaster that nobody wants to address.
           | 
           | There is real data used for insurance premiums and claims
           | payouts but it's being swapped out for AI slop, and the sales
           | folks are getting bonuses for selling hot garbage and the
           | executives are getting bonuses for buying hot garbage.
        
             | ffsm8 wrote:
             | Huh, if sales uses AI/LLMs and succeeds with it, as it
             | seemed to me youve implied with their bonuses ... Isn't
             | that actually positive for the company?
             | 
             | Or do you mean they succeed by promising lies via AI?
        
               | narutosasuke wrote:
               | Not OP but I could see this being short term savings
               | related to the cost of sourcing/generating risk data
               | leading to bonuses prior to the deficiencies in that risk
               | model being exposed in claims long term.
        
               | aprilthird2021 wrote:
               | The industry is insurance. If you offload your due
               | diligence to an AI that's wrong, the company will go
               | underwater. What if the LLM tells you it's found a great
               | market with little competition, high net worth
               | individuals, etc. etc. and you don't check and end up
               | making all your sales in the hills of California in
               | wildfire country?
        
         | sdesol wrote:
         | I would say this is true until we hit a "oh shit, they can
         | really sue us, even if we warn them" moment. I'm imagining
         | legislation will come into play that will make it less ideal
         | for business critical solutions. I know Air Canada has been
         | sued and probably others.
         | 
         | I still think we are in the honeymoon phase, and once that is
         | over, LLM will become what it is meant to be, which is a power
         | tool for domain experts.
        
       | ascorbic wrote:
       | The o1 example is interesting. In the CoT summary it acknowledges
       | that the most recent official information is 1611m, but it then
       | chooses to say 1622 because it's more commonly cited. It's like
       | it over-thinks itself into the wrong answer.
        
         | freehorse wrote:
         | Does it search the internet for that? I assume so because else
         | claiming how often something is cited does not make sense, but
         | would be interesting to know surely. Even gpt4o mini with kagi
         | gets it right with search enabled (and wrong without search
         | enabled - tried over a few times to make sure).
        
           | asl2D wrote:
           | Can the claim about citation frequency be just an answer
           | pattern and not model's exact reasoning?
        
             | freehorse wrote:
             | Yeah there could be parts of the training set with 1611
             | being explicitly called the official and 1622 being
             | explicitly called the most common answer. But it can also
             | have access to search results directly I think. Is there a
             | way to know if it does or not?
        
           | sd9 wrote:
           | I don't think the public o1 can search the internet yet,
           | unlike 4o. In principle it could know that something is more
           | commonly cited based on its training data. But it could also
           | just be hallucinating.
        
             | diggan wrote:
             | > In principle it could know that something is more
             | commonly cited based on its training data
             | 
             | Could it? Without explicit training for that, how would it
             | be expected to know it has to be able to count occurrences
             | of something?
        
               | sd9 wrote:
               | I think it would be more vibes based - commonly occurring
               | things would be reinforced more in the weights. Rather
               | than it explicitly counting the number of occurrences.
        
               | diggan wrote:
               | So the probabilities would be skewed towards something,
               | but unless the model could somehow count/infer its own
               | weights, I don't see how it could "introspect" to see if
               | something is more common than something else.
        
             | freehorse wrote:
             | > it could know that something is more commonly cited based
             | on its training data
             | 
             | No there is no such concept or way to do something like
             | that. LLMs do not have such kind of meta-knowledge over
             | their training data or weights. But there could be explicit
             | mentions about this on their training data and they could
             | pick on that and that is probably the simplest explanation.
        
               | og_kalu wrote:
               | >LLMs do not have such kind of meta-knowledge over their
               | training data or weights.
               | 
               | Not sure this is a claim that can be confidently made.
               | 
               | https://arxiv.org/abs/2309.00667
               | 
               | https://x.com/flowersslop/status/1873115669568311727?t=eB
               | MbK...
        
         | blueflow wrote:
         | How could a language model infer that the official information
         | overrules anything else?
        
           | mistercow wrote:
           | I'm not sure what kind of response you're looking for, or if
           | this is a rhetorical question or not. But "how could a
           | language model infer...?" can be asked about a whole lot of
           | things that language models have no problem reliably
           | inferring.
        
             | blueflow wrote:
             | > that language models have no problem reliably inferring
             | 
             | ... the article did give me a different impression.
        
               | mistercow wrote:
               | I don't think you read my comment correctly.
        
           | ben_w wrote:
           | Same way as we can: learning which sources are more
           | trustworthy.
           | 
           | There's limits to how far you can go with this -- not only do
           | humans make mistakes with this, but even in the abstract
           | theoretical it can never be perfect:
           | https://en.wikipedia.org/wiki/Munchhausen_trilemma -- but it
           | is still the "how".
        
             | fullstackwife wrote:
             | for the last 25+ years we rather not learned, but trusted
             | the top3 of SERPs. Every ranking algorithm will be gamed
             | eventually
        
               | ben_w wrote:
               | I would say that we learned to trust the search engines;
               | but otherwise I agree with you: every ranking algorithm
               | will be gamed eventually.
               | 
               | (I wonder if giving an LLM content with intent to cause
               | its users to spend money they didn't need to, would count
               | as fraud, hacking, both, something else entirely?)
        
           | HeatrayEnjoyer wrote:
           | Attention models learn what to pay attention to.
           | 
           | It's been found that data that begin with "Wikipedia:" are
           | automatically weighted higher by language models during
           | training, completely unsupervised.
        
             | blueflow wrote:
             | But this is the same problem - Wikipedia is a secondary
             | source and should always get overruled by the primary
             | source.
        
               | dotancohen wrote:
               | And there are documented cases of concentrated efforts to
               | manipulate the content of Wikipedia.
        
               | brookst wrote:
               | You're probably being facetious but just in case:
               | 
               | https://www.bbc.com/news/technology-28481876.amp
               | 
               | https://www.bbc.com/news/technology-58559412.amp
               | 
               | (With apologies for amp links)
        
               | dotancohen wrote:
               | I was most certainly not being facetious.
        
         | patrulek wrote:
         | I think i had similar case yesterday for Python script. It gave
         | me code for older version of a module, but when i pasted the
         | error i got, it corrected itself and gave me proper solution
         | for version i had installed.
        
       | Terr_ wrote:
       | > Welcome to the era of generative AI, where a mountain can have
       | multiple heights, but also only one height, and the balance of my
       | bank account gets to determine which one that is. All invisible
       | to the end user and then rationalised away as a coincidence.
       | 
       | I've always found the idea of untraceable, unfixable,
       | unpredictable bugs in software... Offensive. Dirty.
       | Unprofessional.
       | 
       | So the last couple years have been been disconcerting, as a non-
       | trivial portion of people who I _thought_ felt similarly started
       | to overlook it in LLMs, while also integrating those LLMs into
       | flows where the bad-output can 't even be detected.
        
         | choeger wrote:
         | As it turns out, correctness very often simply doesn't matter.
         | Or not as much as one would intuitively think.
         | 
         | How many shops are there optimizing "business strategies" with
         | data that's -essentially- garbage?
        
           | croes wrote:
           | For that LLMs are good but I bet some people want to use it
           | for things where correctness is vital.
        
             | scarface_74 wrote:
             | In that case you use RAG and have it tell you the source.
        
               | dotancohen wrote:
               | A RAG needs to be implemented by the LLM provider. The
               | simple end user has no idea what that means, even though
               | he will be (incorrectly) using the LLM for a vital
               | purpose.
        
               | scarface_74 wrote:
               | ChatGPT does exactly that with its built in runtime and
               | web search.
               | 
               | But the LLM provider doesn't have to do that. Langchain -
               | the Python AI library - and OpenAI's own library has
               | support for third party tools.
               | 
               | It's up to third parties to build on up of it.
        
           | delusional wrote:
           | > How many shops are there optimizing "business strategies"
           | with data that's -essentially- garbage?
           | 
           | How many of those shops are knowingly optimizing with
           | garbage?
           | 
           | I'd argue that most of this data, which I would agree is
           | garbage, is actually processed into seemingly good data
           | through the complex and highly human process of self-
           | deception and lies.
           | 
           | You don't tell the boss that the system you worked 2 month on
           | is generating garbage, because then he'll replace your with
           | someone who wouldn't tell him that. Instead you skirt
           | evaluating it, even though you know better, and tell him that
           | it's working fine. If the idiot chooses to do something
           | stupid with your bad data, then that's his problem.
        
       | firtoz wrote:
       | Looking towards the future, we will need to move away from
       | "tokens are characters to print", We're kind of starting to
       | consider this with "tool calls" but I believe an architectural
       | shift will become necessary.
       | 
       | We do have some kind of understanding of what kind of concept we
       | want to emit next, e.g.
       | 
       | ```
       | 
       | [The height:property name] of [Mount Bartle Frere:proper noun,
       | describing an object to get a property out of], [in
       | metres:attributes], is [?: retrieve value | (Mount Bartle
       | Frere).("height", "metres")].
       | 
       | ```
        
         | giantrobot wrote:
         | Uh oh, that sounds suspiciously like querying structured data.
         | You can't hype SQL or worse, SPARQL, to investors!
        
           | firtoz wrote:
           | Well, you can hype the results and then put it into a black
           | box and call it a LLM anyway
           | 
           | Which is pretty much what O1 etc are
           | 
           | Update: it seems your recent submission[1] is pretty much
           | that... interesting :D
           | 
           | 1: https://github.com/caesarhq/textSQL
        
         | DoctorOetker wrote:
         | Are there any minimal / micro LLM's that are paired with a
         | large RAG vector database, but still have performance on par
         | with the huge LLM's? i.e. shifting parameters from weights to
         | vector database, so that a smaller LLM can fit in RAM, and the
         | vector database on disk. Possibly multiple calls.
         | 
         | When you ask a human to switch context (changing topic) or to
         | change activity (e.g. football to table tennis), they typically
         | need some warm-up too, so it seems excessive to have all
         | knowledge in high bandwidth RAM.
         | 
         | It would seem basic mathematics, set theory etc should stay in
         | RAM.
        
       | tucnak wrote:
       | I don't get it why people demo COT reasoning with o1 when there's
       | models like Gemini 2.0 Thinking that would usually solve the same
       | tasks and would happily produce the full output.
        
       | stereobit wrote:
       | Is there no concept like page rank that biases certain inputs to
       | have higher impact while training based on recency and
       | popularity?
        
       | scarface_74 wrote:
       | This problem is simple to solve for most real world use cases.
       | Don't trust any facts from an LLM and use your own trusted source
       | of information and RAG where it will give you citations
       | 
       | https://chatgpt.com/share/6783df4c-904c-8010-a4b5-7301faea3b...
       | 
       | https://chatgpt.com/share/6783e0b8-ce78-8010-9177-d95eb77eac...
       | 
       | I use NotebookLM for most of my real world work these days with
       | my project documentation.
       | 
       | Our company standard is GSuite and NotebookLM is specifically
       | allowed.
        
       | Workaccount2 wrote:
       | How would we rule out that the model didn't notice that the
       | difference was small, and then simply put less weight into
       | determining which is true.
       | 
       | I get the authors point, but I would have liked to see and
       | example with a more egregious error.
        
       | throwawaymaths wrote:
       | kind of crazy that models moving forward don't just strip all
       | multi-numeral tokens. Would be great for llm providers, too,
       | since their tokens consumed will go up.
        
       | antirez wrote:
       | Prompting Claude to show the ambiguity:
       | 
       | Tell me the height of Mountain Bartle Frere. Please don't output
       | any long text, also don't output a single height if you saw
       | multiple heights around. Give me a list of potential heights
       | cited around.
       | 
       | LLM:
       | 
       | Mount Bartle Frere in Queensland, Australia has commonly cited
       | heights of:
       | 
       | 1,622 meters (5,322 feet)
       | 
       | 1,611 meters (5,285 feet)
       | 
       | Since this is quite specific geographic information that may
       | appear in only a few sources, I should note that I may
       | hallucinate details - you should verify these numbers.
        
         | sdesol wrote:
         | I asked a bunch of LLMs using your prompt. The only change was
         | to just use meters.
         | 
         | https://beta.gitsense.com/?chat=bb57a248-e14a-4f33-bbe9-2fa9...
         | 
         | 1622m is most agreed upon. The interesting numbers are the ones
         | with less than 50% agreement. Not sure if they are
         | hallucinations or if they are outdated data.
         | 
         | Click the conversation link in the user message bubble to see
         | the response from each LLM.
        
       ___________________________________________________________________
       (page generated 2025-01-12 23:02 UTC)