httpjames

How well can LLMs solve chess puzzles?

https://sh.itjust.works/post/16775965

How well can LLMs solve chess puzzles? - sh.itjust.works

Each LLM is given the same 1000 chess puzzles to solve. See puzzles.csv. Benchmarked on Mar 25, 2024. | Model | Solved | Solved % | Illegal Moves | Illegal Moves % | Adjusted Elo | |--------------------------|--------|----------|---------------|-----------------|--------------| | gpt-4-turbo-preview | 229 | 22.9% | 163 | 16.3% | 1144 | | gpt-4 | 195 | 19.5% | 183 | 18.3% | 1047 | | claude-3-opus-20240229 | 72 | 7.2% | 464 | 46.4% | 521 | | claude-3-haiku-20240307 | 38 | 3.8% | 590 | 59.0% | 363 | | claude-3-sonnet-20240229 | 23 | 2.3% | 663 | 66.3% | 286 | | gpt-3.5-turbo | 23 | 2.3% | 683 | 68.3% | 269 | | claude-instant-1.2 | 10 | 1.0% | 707 | 66.3% | 245 | | mistral-large-latest | 4 | 0.4% | 813 | 81.3% | 149 | | mixtral-8x7b | 9 | 0.9% | 832 | 83.2% | 136 | | gemini-1.5-pro-latest* | FAIL | - | - | - | - | Published by the CEO of Kagi!

Mar 25, 2024 at 5:13pmWeb
Show threadBlóðbókMar 25, 2024
Not well, apparently.
Show threadconciselyverboseMar 25, 2024
I wonder how many of the ones they “solved” were just because they’d seen it discussed, considering the puzzles are apparently from a public resource.
Show threadBlóðbókMar 25, 2024
Yeah, I don’t know why anyone knowledgeable would expect them to be good at chess. LLMs don’t generalise, reason or spot patterns, so unless they read a chess book where the problems came from…
Show threadCarroladeMar 25, 2024

Likely close to 100%. If you read the (rather good) article, a little further down they test whether the LLM can play an extremely simplistic “Connect 4” game they devise, as a way of narrowing down on specifically reasoning capabilities.

It cannot.

Chess puzzles, in particular, are frequently shared and discussed in online chess spaces, so the LLM will have a significant amount of material to work with when it tries to predict the best response to give to the prompt.

Show threadColeSlothMar 26, 2024
I didn’t figure. I’m sure they could be taught to be much better, but normal computing can already play chess more or less perfectly. There really isn’t much any room left to be gained.
Show threadAFK BRB ChocolateMar 25, 2024
People thinking LLMs should be even serviceable at chess didn’t understand LLMs. They really aren’t problem solving applications. They’re optimized for making responses to questions that look like what a response should look like, not for being accurate. That’s really clear if you ask them for mathematical proofs. They will generate proofs that look like the right sort of thing, but they won’t be correct unless they have the specific proof in their training data.
Show threadsnaggenMar 26, 2024
This is obvious for people who understand the basics of LLM. However, people are fooled by how intelligent these LLM sounds, so they mistake it for actually being intelligent. So, even if this is an open door, I still think it’s good someone is kicking it in to make it clear that llms are not generally intelligent.
Show threadAFK BRB ChocolateMar 26, 2024
Agreed, it’s good to have these kinds of articles so people get a better feel for what tools like this are and aren’t.
Show threadlatetolemmyMar 26, 2024
You got the self confidently ignorant redditor tone down pat. But if you had read a few papers on llm you would realize that you’re wrong.
Show threadscorpiousMar 26, 2024
Are you going to share any of your wisdom here? Or rebut misinformation?
Show threadDocusMar 25, 2024

Of course they are bad at solving problems. The I in LLM stands for intelligence.

(Credit: …haxx.se/…/the-i-in-llm-stands-for-intelligence/)

The I in LLM stands for intelligence

I have held back on writing anything about AI or how we (not) use AI for development in the curl factory. Now I can't hold back anymore. Let me show you the most significant effect of AI on curl as of today - with examples. Bug Bounty Having a bug bounty means that we offer … Continue reading The I in LLM stands for intelligence →

daniel.haxx.se
Show threadlatetolemmyMar 26, 2024
Guess your name doesn’t have an I as well
Show threadBetaDoggo_Mar 25, 2024
This has more to do with how much chess data was fed into the model than any kind of reasoning ability. A 50M model can learn to play at 1500 elo with enough training: adamkarvonen.github.io/…/chess-world-models.html
Chess-GPT’s Internal World Model

A Chess-GPT Linear Emergent World Representation

Adam Karvonen
Show threadGeneral_EffortMar 26, 2024
I wonder why gpt-4 is so good at chess.
Show threadbionicjoeyMar 26, 2024
If I tried to make an illegal move 20% of the time, would you also say I am good at chess?
Show threadGeneral_EffortMar 26, 2024
Depends on circumstances, obviously.
Show threadbionicjoeyMar 26, 2024
Okay. What if the circumstance is because I’m just recalling a bunch of chess puzzle solutions I’ve seen before and regurgitating the one I think is the correct solution for this particular pizzle without really understanding the rules of chess?
Show threadGeneral_EffortMar 26, 2024
That’s another thing I’m wondering about, but so is anyone. I’d still want to know why GPT-4 does so much better than the others.
Show threadnyanMar 26, 2024
I’m actually a bit surprised they got any of them right. Maybe the ones they solved correctly had exact matches in their training data . . . ?
Show threadBuffaloxMar 27, 2024
This to me shows that LLM simply isn’t trustworthy, it’s one thing it can’t solve a puzzle, fair enough. But that it uses illegal moves is kind of alarming.
This is a relatively simple task, so this proves that LLM isn’t trustworthy even for simple tasks.
That said, I still think LLM is an impressive technology, but I’d be very careful relying on it for anything. The fact that some companies already use them for customer support gives me horror goosebumps.
Show threadCyyyMar 28, 2024

i mean… try using it for even simple stuff like designing code. Often, ChatGPT creates a fantasy library that does the task you ask chatgpt tk do… the library don’t exists, but chatgpt writes you code using that fantasy software library. Same with program functions who don’t exist.

Same happens with stuff like People, telephone numbers, locations, books etc… tons of fantasy stuff.

LLMs aren’t trustworthy for such stuff if you need real info and not just creative help with fantasy stuff. And even for those tasks it is usually not really good enough.

Show threadBuffaloxMar 28, 2024
Thanks, I wasn’t aware of that, the “fantasy” stuff was news to me when I read they make illegal moves to solve chess puzzles.
There has been a lot of praise for how amazing these LLM models are, but apparently they have some pretty serious limitations, that might even make them dangerous to use, if you are not aware.
Show threadCyyyMar 28, 2024
The issue with LLMs is, that they got trained with all kinds of data. so not just real scientific data but also fantasy (lies, books, movie scripts etc )… and nobody told the LLMs while training them what is fantasy or what not. so they only know how to generate text that looks “legit” without really knowing what is true and what not. so if you ask for a person and their personal details as an example… a LLM could generate real looking data that is just fantasy because it learned that such data looks like this.
Show threadBuffaloxMar 28, 2024
I’m sure you are right, LLM isn’t intelligent enough to distinguish between fact and fantasy on it’s own, which IMO is a bit disappointing considering early reports about ChatGPT. Which were overwhelmingly positive. The AI is way more artificial than intelligent. Or as I saw earlier, the i in LLM is for intelligence. 😋