Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Benchmarking GPT-4 Turbo – A Cautionary Tale

blog.mentat.ai

Post reply on HN
Print view View original

31–40 of 118 posts

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#31

Post by boxed »

famouswaffles wrote: post #23
boxed wrote: post #20

Earlier quoted context omitted.

When people say "oh look how amazing, it can solve programming problems!" when in fact it has only seen the models CHEAT, is an enormous problem. For cases where finding the answer it's perfectly fine, but it's not fine for claims that it can code. There's a huge difference.

Okay... Funny how forcing it to not CHEAT did not increase apparent ability. It can code and it has memorized some coding questions are not mutually exclusive.

> Okay... Funny how forcing it to not CHEAT did not increase apparent ability.

The article did the opposite. It forced the models to cheat to solve the problems. Which it did happily. It should have stated "there is no actual problem to solve here, you must supply a problem for me to solve".

> It can code and it has memorized some coding questions are not mutually exclusive

This I will give you. Many humans try to cheat at basic math because they are lazy, so will this model. Maybe that's a sign of intelligence :P

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#32

Post by GaggiX »

boxed wrote: post #4

> We designed a test for this theory: we reran the benchmarks without showing the models the instructions to each exercise. Instead, we just told them that they were Exercism exercises, and gave them the exercise names and function stubs. This summarizes all my skepticism agains the AI field. Pretty clear that they aren't solving the problem, they have them memorized.

So how can it solve novel problems? Internet does not have all combinations for every possible task with any random programming language, library or constraints. It can even solve problems with non-existing programming languages and libraries if you describe them, if that's just memorization then I don't know what it isn't.

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#33

Post by ConnorMooneyhan »

FeepingCreature wrote: post #26

I wonder how often a human could guess the exercise based on just the function stub.

yeah, some of the exercises are like the following:

```

function helloWorld() {

  return "";
}

helloWorld()

```

but those sorts of obvious examples are mostly in the beginner exercises, so I wonder what the distribution of the correct answers was. If it was guessing based on function stubs, the prediction would be that correct answers would be clustered around the beginner exercises, and that as the exercises advanced in difficulty, there were fewer correct answers.

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#35

Post by phillipcarter »

msp26 wrote: post #25

I'm interested in more testing on the context side of things. For my NLP pipelines, I batch n-articles together to process (extract fields from) in one prompt (final output is something like this {"1":[{}], "2": [{},{}]...}) in one message. Compute-wise it's inefficient but OpenAI charges by the token so it doesn't matter. It's very reliable on gpt-4 8k. I was also pretty happy with the results on 4-turbo initially b…

Interesting. I was skeptical about some of their claims regarding longer context, since it's been my experience that these models just get lost after enough of it.

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#36

Post by hanselot »

famouswaffles wrote: post #23
boxed wrote: post #20

Earlier quoted context omitted.

When people say "oh look how amazing, it can solve programming problems!" when in fact it has only seen the models CHEAT, is an enormous problem. For cases where finding the answer it's perfectly fine, but it's not fine for claims that it can code. There's a huge difference.

Okay... Funny how forcing it to not CHEAT did not increase apparent ability. It can code and it has memorized some coding questions are not mutually exclusive.

Though this is exactly what happened. The initial test was ran on a model that "Cheated" (aka has memorized the answers). The second test was run on a model that didn't "Cheat" as much, yet still got only 2% less score. So, the question is not resolved really. How much did the first model cheat, and how much did the second? If the second model "cheats" less, then it wins.

Also, I don't understand your obsession with the word cheating. If you have solved a problem before on a different website and solve it again, did you cheat? Or did you just use your brain to store the solution for later?

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#37

Post by maciejgryka »

I have similar conclusions so far. We have a custom data set (basically visual Q&A about web apps) and `gpt4` gets roughly 90% correct, while `gpt-4-1106-preview` only 86%. It's a little noisy (I didn't yet check out the new seeds functionality), but roughly consistent.

Since I created this dataset by hand, it can't really be memorized. I'm sure there's _similar_ data in the training set, but answering correctly still requires some reasoning-like capabilities.

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#38

Post by raincole »

boxed wrote: post #31
famouswaffles wrote: post #23

Earlier quoted context omitted.

Okay... Funny how forcing it to not CHEAT did not increase apparent ability. It can code and it has memorized some coding questions are not mutually exclusive.

> Okay... Funny how forcing it to not CHEAT did not increase apparent ability. The article did the opposite. It forced the models to cheat to solve the problems. Which it did happily. It should have stated "there is no actual problem to solve here, you must supply a problem for me to solve". > It can code and it has memorized some coding questions are not mutually exclusive This I will give you. Many humans try to ch…

Me: What's 6x6?

You: 36

Me: You cheated! You just cited the answer you memorized! You should have started from addition.

You: ...okay? 6+6=12, 12+6=18, 18+...

Me: You cheated again! You just have 6+6=12 memorized! You should make the rule of addition out of Peano axioms.

You: ...you're being annoying, but okay? First axiom, we define 0 as...

Me: You cheated again! You memorized Peano Axioms! Jesus Christ, is there any intelligent creature left?

Re: Benchmarking GPT-4 Turbo – A Cautionary Tale

#40

Post by broast »

m3kw9 wrote: post #39

Now do a programming task that requires more than 32k of context and see who’s “better”. If you don’t bench mark that you cannot get an overall pic. GitHub copilot for example could benefit big from the increased context

Obviously it's a drawback but the silver lining of the small context window is it forces me to decouple everything and have very sensible and strict api's where I just write the docs and it writes the code.
Post reply on HN