How outdated information hides in LLM token generation probabilities

https://blog.anj.ai/2025/01/llm-token-generation-probabilities.html

by anjsimmo • 2 years ago

123 53 2 years ago

blog.anj.ai

0xKelsey• 2 years ago

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.

aprilthird20212• 2 years ago

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• 2 years ago

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.

sdesol2• 2 years ago

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• 2 years ago

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.

freehorse3• 2 years ago

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• 2 years ago

Can the claim about citation frequency be just an answer pattern and not model's exact reasoning?

sd92• 2 years ago

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.

blueflow3• 2 years ago

How could a language model infer that the official information overrules anything else?

mistercow2• 2 years ago

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.

ben_w2• 2 years ago

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/Münchhausen_trilemma — but it is still the "how".

HeatrayEnjoyer2• 2 years ago

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.

patrulek• 2 years ago

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_3• 2 years ago

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• 2 years ago

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• 2 years ago

For that LLMs are good but I bet some people want to use it for things where correctness is vital.

delusional• 2 years ago

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.

firtoz2• 2 years ago

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• 2 years ago

Uh oh, that sounds suspiciously like querying structured data. You can't hype SQL or worse, SPARQL, to investors!

firtoz2• 2 years ago

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

Terr_3• 2 years ago

A little bit RDF-y too.

https://en.wikipedia.org/wiki/Resource_Description_Framework

DoctorOetker• 2 years ago

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• 2 years ago

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• 2 years ago

Is there no concept like page rank that biases certain inputs to have higher impact while training based on recency and popularity?

scarface_743• 2 years ago

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• 2 years ago

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• 2 years ago

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• 2 years ago

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.

sdesol2• 2 years ago

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.

cyanydeez• 2 years ago

Obviously, models need the same brain flush brains do when they dream

gwern• 2 years ago

The bank account anchoring bias may be just the base model predicting as best as it can, as OP described earlier. Imagine a random Internet page where you read "There is V in my X. The Y of Z A, in Bs, is _." What do you expect? V, of course! Because why would a person on the Internet ever join these two factoids if they were totally unrelated (as they are in fact in this example)? That would be a gross violation of basic writing and human communication norms (https://en.wikipedia.org/wiki/Cooperative_principle).

How outdated information hides in LLM token generation probabilities

Related Stories