3 ms·
> a complex pattern-fitting supercomputer hovering over a massive table of precomputed patterns That was perhaps true of earlier and smaller LLMs, like GPT-1 a
by fl7305 3y ago
> a complex pattern-fitting supercomputer hovering over a massive table of precomputed patterns
That was perhaps true of earlier and smaller LLMs, like GPT-1 and GPT-2.
But as they grew larger and were trained with more and more data, they changed from pure pattern matching to implementing algorithms to compress more information into their structure than pure pattern matching can achieve.
These algorithms are incomplete and buggy, but they are nonetheless executing algorithms, and not just pattern matching.
This phenomenom can be seen in toy-sized neural networks. For instance, addition of two input values modulo a constant. As a small network is trained, at some point the internal structure can change from pattern matching to implementing addition using Fourier transforms. This is clearly visible in its structure. The network now performs the task perfectly for all inputs, regardless of having seen them in training.
You can ask ChatGPT 4 to execute an algorithm for you. I just tried this one:
I would like to play a game, where you are the host. We start off with a score that is 1234143143. At the start of each turn, you tell me the current score and ask me if I want to play a or b. If I choose a, the score is halved, and 30 is added. If I choose b, the score is doubled, and 40 is subtracted. Only use integers and round down.
It will happily execute this algorithm. For large numbers, it is slightly off on the arithmetic. When I asked it to double check, it did so using Python code. After that, it kept using Python code to perform the math. It was also able to reason intelligently about different outcomes if always picking a (or b) given different starting points.
Now, if you have enough memory and training data, of course you can build a gigantic lookup table that has this exact text sequence in it to replicate "executing the algorithm" I described.
Is that your claim? How much memory are we talking about? My feeling is that it'd be far more than the number of atoms in the universe.
PS
Me: How far it is to the Simpsons house in Springfield?
ChatGPT: The Simpsons' house in Springfield is a fictional location from the animated TV series "The Simpsons." Since Springfield is a fictional town and its location is not consistently specified in the series, it's not possible to determine a real-world distance to the Simpsons' house.
Me: Do I have a brother?
Please answer with either of:
a) The answer
b) There is an answer, but I do not know
c) There is no answer
ChatGPT: b) There is an answer, but I do not know
- andsoitis 3y ago> It will happily execute this algorithm. For large numbers, it is slightly off on the arithmetic. When I asked it to double check, it did so using Python code. After that, it kept using Python code to perform the math. It was also able to reason intelligently about different outcomes if always picking a (or b) given different starting points. Notice that you had to notice the error and had to prompt it to double check. Lots of complicated things going on here. Many (most?) humans will fail somewhere along this trajectory. Did it double check the Python code to make sure it is correct (not just in the sense that it is valid, executable code, but that it is the correct check in the first place)? Or did you double check that its modified algorithm is correct? Fool me once and all that… Upon reflection it appears as if you have a heuristic (algorithm? that leverages logic, awareness, critical thinking, experience, a goal in mind, intuition, etc. to push towards better results. “It was able to reason intelligently” imbues qualities that I am skeptical is reasonable to attribute to this very narrow domain - what’s an example where it showed intelligent reasoning capabilities?
- fl7305 3y ago> Notice that you had to notice the error and had to prompt it to double check. Lots of complicated things going on here. Many (most?) humans will fail somewhere along this trajectory. Sure. This was covered by my statement above: "These algorithms are incomplete and buggy". > “It was able to reason intelligently” imbues qualities that I am skeptical is reasonable to attribute to this very narrow domain - what’s an example where it showed intelligent reasoning capabilities? Here's an example. I asked it to analyze the case where we always pick option "b" in my example above. It took my word problem and boiled it down to an equation: ChatGPT: To find the threshold where the behavior switches from decreasing to increasing when always choosing option "b", we need to identify the smallest integer starting value that, after being doubled and reduced by 40, leads to an equal or larger integer in the next iteration. We're looking for the smallest integer x where 2x - 40 >= x This was part of a longer conversation where it analyzed different properties and outcomes of the "game rules" that I gave it. As you pointed out, it got some things wrong and had to be corrected. But Socratic reasoning works fairly well to guide it. It can find errors in its own reasoning. For instance, if asked to actually calculate a few iterations for a given case, it will find its own errors in its claims about that case. Is it useful right now? Maybe, maybe not, depends on your use case. It definitely takes a lot of thinking on your own and guiding it. At some points it goes from seemingly intelligent to downright pigheaded and stupid. But in my view there is absolutely no way a lookup table algorithm can contain enough data to be anywhere near the level of responses we're seeing here.