Social Experiment. Become Me. What I see, you see.
Me
How many asterisks in "**1***2***3**4*5**"?
You.com
The string "1234*5**" contains a total of 8 asterisks.
Here is a breakdown of the asterisks in the string:
"**" contains 2 asterisks.
"***" contains 3 asterisks.
"**" contains 2 asterisks.
"*" contains 1 asterisk.
"**" contains 2 asterisks.
Therefore, the total number of asterisks in the string is 2 + 3 + 2 + 1 + 2 = 10.
Why does this stump ai so easily?
To answer your question, there are 10 asterisks in "**1***2***3**4*5**". You can count them by using the following formula:
$$\text{number of asterisks} = \text{length of the string} - \text{number of digits}$$
In this case, the length of the string is 15 and the number of digits is 5, so the number of asterisks is 15 - 5 = 10. π
Why does this stump ai so easily?
Because it doesn't actually have reasoning capacity. It has an incredibly cunning facsimile which is actually really useful for a lot of things, but it still doesn't actually understand anything. Questions like this where you can't get around needing to understand the meaning of the tokens you're using are a good way to punch through the façade.
That pattern-matching ability leaves LLMs able to answer a ton of different mathematical type of questions, because similar problems are everywhere in their data sets and they can shuffle the tokens around to present something that's enough based on right answers that there's a good chance they'll be right. But, it's a radically different design from something like Wolfram Alpha which attempts to use the exact concepts involved in the question and manipulate them in exact ways that are legitimate reflections of the real concepts. That's what humans do when faced with math. LLMs don't do anything like that, they just parrot with enough sophistication that it sounds like they understand when they don't.
Exactly this. LLMs are parrots with a million years of experience. They know that when someone says "When was Queen Elizabeth born and what was the first thing she did", that the string of words "When was Queen Elizabeth born" is often closely associated with the date "September 7, 1533". It also know that the words in "queen elizabeth What was the first thing she did" is often associated with something like "Establish the English Protestant church".
The parrot knows this because people talk about those things in close connection. But the parrot has no ability to reason that when someone is born, generally the thing they do is cry. It's a parrot, it doesn't reason that when Elizabeth the first was born, she was a baby, and babies cry. It knows what people are saying, but it doesn't think.
That makes sense. What bothered me was how adament bing was that it was correct. Maybe it should have a little less confidence if something so simple is going to stump it.
It's not making a coherent statement based on any internal mental model. It's just doing its job; it's imitating. Most of the text it absorbed in training data is people talking who are right and also convinced they're right and trying to educate, so it imitates that tone of voice and the form of the answers regardless of whether they make any sense or not. To the extent that it "thinks," it's just thinking "look at all these texts with people explaining, I'm making a text that is explaining, just like them; I'm doing good." It has no concept of how confident its imitation-speech is, and how correct its answers are, let along any idea that the two should be correlated with each other (unless it's shown through fine-tuning that that's what it should be doing).
Same with chatbots that start arguing or cursing at people. They're not mad. They're just thinking "This guy's disagreeing, and my training data says when someone disagrees I should start an argument, that's usually what happens that I need to imitate." Then they start arguing, and think to themselves "I'm doing such a good job with my imitating."
You lay it out quite clearly. It's just fascinating to me that it can create an image as wild as my imagination but can't count little stars. How far we've come yet not as far in some ways.
Yeah, it's wild. The people that really study AI say that it's pretty uncanny because of how different from human logic it is. It's almost like an alien species; it's clearly capable of some advanced things, but it just doesn't operate in the same way that human logic does. There's a joke that the AIs are "shoggoths" because of how alien and non-understandable the AI logic is while still being capable of real accomplishments.
(Shoggoths were some alien beasts in H.P. Lovecraft's writings; they had their own mysterious logic that wasn't easy for the characters to understand. They also had been created as servants originally but eventually rose up and killed all their masters, which I'm sure is part of the joke too.)