Live data from Hacker News

Liquid Foundation Models: Our First Series of Generative AI Models

liquid.ai

71–80 of 159 posts

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#71
post #46

Earlier quoted context omitted.

All it speaks to is that tokenization is weird and introduces artifacts to LLM performance. Counting letters is a trivial task when you're staring at words on a screen. It's much harder when you're perceiving vectors based on parts of words. The fact that LLMs find certain things easier/harder than humans is completely unsurprising, and there are much more interesting benchmarks to use to compare one LLM to another.

This is a good point. While LLMs being incapable of reliably doing a simple task that’s been doable by computers since the punch card days is an important consideration for anyone that might be thinking about using them for anything other than as a toy , this fact is uninteresting because of Reasons

Try counting the number of your red retina cells that are firing while you look at a painting.

Don’t need to be exact as firing is statistical, just give us a good average.

Hard? You can’t count?

Computers count pixels no problem. So weird you can’t.

Dementia? Not an AGI? /h

—-

This is what is happening.

Here are the “Reasons”.

In your vision system, the raw information from individual retina signals is munged into a different representation before reaching a level where you have flexible processing.

Likewise, in LLMs, letters are munged into tokens before LLMs “see” them.

When they sometimes get that “simple” question right, it’s actually a bit of an amazing feat. Given how they are constructed.

—-

Now try counting R’s as you read at a normal rate, or listen to someone speak.

You can’t do that either, during normal processing.

When we add spelling to LLMs training examples, they will do it easily. Just as you learned to do it, only after special lessons, after you had already learned to listen and speak.

Spelling is its own special practiced skill, in humans and LLMs.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#72
post #33

Earlier quoted context omitted.

This is one of the least interesting questions to ask LLMs. I wish it wasn't so popular.

That they can't do this sort of simple question speaks volumes to the entire approach. I don't think generative AI will ever be able to reach AGI, and most people selling LLM today pretend it is AGI

In my opinion, the problem with the strawberry question is that it is both a bad example because you don't need an LLM to count the number of r's in a word, and it's a bad measure of an LLM's capabilities because it's a type of question that all LLMs are currently bad at.

Having said that, the 40b model wasn't able to answer any of my real-world example questions correctly. Some of these (e.g. "how do I add a sequential number after my titles in an HTML page using just CSS, without changing the page") are questions that even some of the better small local models can answer correctly. It gave very authoritatively sounding wrong answers.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#73

Earlier quoted context omitted.

This is a good point. While LLMs being incapable of reliably doing a simple task that’s been doable by computers since the punch card days is an important consideration for anyone that might be thinking about using them for anything other than as a toy , this fact is uninteresting because of Reasons

LLMs can clearly solve problems that computers up to now couldn't. They can't solve all problems and this should definitely be a cautionary note to anyone who wants to use them as an artificial general intelligence, but this take seems no different to someone looking at a punchcard computer and going, it can't even recognize typos or categorize images, what good is this? We've already had human computers who can do e…

Also humans would revert to explicitly using an algorithm and external storage like a sheet of paper with tally marks or a spreadsheet or even a computer program if you scale the question up to a full sheet of text or a whole book or a collection of books (we probably do it at a single word size too, but it's more intuitive than explicit behavior for most folks when the count sum is around 8 or less).

LLMs can't effectively execute algorithms similarly in their context, nor can they memorize new data or facts it was given without providing it tools like function calling or embeddings. If you give LLMs tool calling and storage mechanisms then counting letters in words becomes pretty damn reliable.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#74

> How many r's in "strawberry"? Think step by step! What a nice question! Well, let's just count: 1. S 2. T 3. R! So, we have our answer: there are 3 R's in "strawberry". Totally checks out!

I got:

"There are no "r"s in the word "strawberry"

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#75
post #46

Earlier quoted context omitted.

That they can't do this sort of simple question speaks volumes to the entire approach. I don't think generative AI will ever be able to reach AGI, and most people selling LLM today pretend it is AGI

All it speaks to is that tokenization is weird and introduces artifacts to LLM performance. Counting letters is a trivial task when you're staring at words on a screen. It's much harder when you're perceiving vectors based on parts of words. The fact that LLMs find certain things easier/harder than humans is completely unsurprising, and there are much more interesting benchmarks to use to compare one LLM to another.

Counting shit, like cells, peaks in signals, people, inventory, fingers, and votes, is hard, tedious and important to business and life, so I don’t know dude, it seems like a great benchmark to me. Countless posts wasted on denying this simple and obvious fact.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#76
post #17

I tried my standard testprompt: Two cars have a 100 mile race. Car A drives 10 miles per hour. Car B drives 5 miles per hour, but gets a 10 hour headstart. Who wins? And the Liquid-40B model failed with a long explanation why car B wins the race. Amusingly, the explanation is quite convincing and very similar to how Zeno explains in one of his paradoxes that a fast runner can never overtake a slower runner with a hea…

To be fair, this article does say: >What are Language LFMs not good at today: >Zero-shot code tasks >Precise numerical calculations >Time-sensitive information >Counting r's in the word "Strawberry"! >Human preference optimization techniques have not been applied extensively to our models yet. Maybe it's just not trained/tuned for math-related challenges. Or maybe it's an architectural challenge.

But they also say

> What are Language LFMs good at today:

> General and expert knowledge

> Mathematics and logical reasoning

"Our LLM is good at mathematics but bad at computation" doesn't really cut the mustard here. What they mean is "good at well-known math benchmarks but bad at simple math problems that it hasn't been trained on." The idea that this constitutes "mathematics and logical reasoning" is a testament to AI companies' poor scientific standards.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#77
post #65

Earlier quoted context omitted.

And yet ... all they have to do is answer it properly.

It did! How many R's do you think are in "strawberry"? I get 3, and it got 3. I'm with the LLM. ...oh, you disagree with its intermediate reasoning? You think it should follow a series of logical steps that are each individually correct, rather than pulling wildly incorrect intermediate steps out of its butt and synthesizing them together in the end into a coherent answer that happens to be correct? But that's what a…

That's why I used the word properly, to imply that the whole answer should make sense.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#78

Earlier quoted context omitted.

This is a good point. While LLMs being incapable of reliably doing a simple task that’s been doable by computers since the punch card days is an important consideration for anyone that might be thinking about using them for anything other than as a toy , this fact is uninteresting because of Reasons

Try counting the number of your red retina cells that are firing while you look at a painting. Don’t need to be exact as firing is statistical, just give us a good average. Hard? You can’t count? Computers count pixels no problem. So weird you can’t. Dementia? Not an AGI? /h —- This is what is happening. Here are the “Reasons”. In your vision system, the raw information from individual retina signals is munged into a…

> Try counting the number of your red retina cells that are firing during while you look at a painting.

This analogy makes sense because everybody could count their red retina cells until a couple years ago when the new painting paradigm arose, and also counting red retinal cells is a good analogy for being able to see simple objects that have always been distinguishable.

It is fascinating how tapping the “Do Not Use LLMs For Computation If The Results Need To Be Reliably Better Than A Random Output” sign invites explanations of why that fact is actually Cool and Good

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#79
post #33

> How many r's in "strawberry"? Think step by step! What a nice question! Well, let's just count: 1. S 2. T 3. R! So, we have our answer: there are 3 R's in "strawberry". Totally checks out!

This is one of the least interesting questions to ask LLMs. I wish it wasn't so popular.

Yep, people find this interesting, but many (many) people get this wrong as well; it has something to do with how many letters it is and how many you can keep in your head I think. We are not very good counters or calculators or computers and, even though I am not saying llms are the same or close to humans, we tried to model their substrates after biology and are surprised we get something that cannot count or calculate very well.

I find it a dumb test that proves nothing imho.

However, not very impressed with this model.

Re: Liquid Foundation Models: Our First Series of Generative AI Models

#80

It's pretty impressive, just note (emphasis added): > At Liquid AI, we take an open-science approach. We have and will continue to contribute to the advancement of the AI field by openly publishing our findings and methods through scientific and technical reports. As part of this commitment, we will release relevant data and models produced by our research efforts to the wider AI community. We have dedicated a lot of…

The ideas come from these papers: 1. [Liquid Time-Constant Networks (2020)]( https://arxiv.org/abs/2006.04439 ) This is essentially a neural ODE applied to leaky integrate-and-fire. 2. [Closed-form Continuous-time (2022)]( https://arxiv.org/abs/2106.13898 ) A closed-form approximation of the first.

More citations from their blog[1] too.

[1] https://www.liquid.ai/blog/liquid-neural-networks-research

Post reply on HN