the model's ablility in math is impressive!
it can solve 3276775*7372793 ,that's incrediable
I do confirm the Q4KM is able to do so as well, using 2 different methods and it does double check.
Therefore, the answer is 24,172,769,175.
It is certainly beyond the average Math^
I do confirm the Q4KM is able to do so as well, using 2 different methods and it does double check.
Therefore, the answer is 24,172,769,175.It is certainly beyond the average Math^
Is that a joke?
I wonder who is the joke here. I am sure it would make a killing with your friends at the troll faculty of science, but for your notice, the percentage of Humans able to give an accurate answer to that using a pen and paper is small, and way fewer could compute that by head. Yes you could use a $5 calculator and use 0.01% of the power of a CR2032, just as you could as well watch HD porn by typing 8008 on these devices.
Anyway, let's use trolls as support for further testings:
Among all 8B models as of March 2025,
55844*2448548
llama3.1: Failure: affirms incorrect answer
Nemotron: Partial success, Answer Ok, but does not affirm (infinite loops)
Gemma: failure: affirms incorrect answer
Olmo: Failure: affirms incorrect answer
GWQ: Failure: affirms incorrect answer
Deepseek: Failure: incorrect answer, but acknowledge
Mistral: Partial Success: Give accurate close approximation, and acknoledge the answer being approximate.
OpenCoder: Partial success: no answers but provided accurate syntax error free C++ code for computation.
OlympicCoder: Failure, provided incorect C++ code and no answers.
phi3: Failure: affirms incorrect answer
I wonder who is the joke here. I am sure it would make a killing with your friends at the troll faculty of science, but for your notice, the percentage of Humans able to give an accurate answer to that using a pen and paper is small, and way fewer could compute that by head. Yes you could use a $5 calculator and use 0.01% of the power of a CR2032, just as you could as well watch HD porn by typing 8008 on these devices.
Anyway, let's use trolls as support for further testings:
Among all 8B models as of March 2025,
55844*2448548llama3.1: Failure: affirms incorrect answer
Nemotron: Partial success, Answer Ok, but does not affirm (infinite loops)
Gemma: failure: affirms incorrect answer
Olmo: Failure: affirms incorrect answer
GWQ: Failure: affirms incorrect answer
Deepseek: Failure: incorrect answer, but acknowledge
Mistral: Partial Success: Give accurate close approximation, and acknoledge the answer being approximate.
OpenCoder: Partial success: no answers but provided accurate syntax error free C++ code for computation.
OlympicCoder: Failure, provided incorect C++ code and no answers.
phi3: Failure: affirms incorrect answer
It's really interesting that you pointed out that instead of asking the AI a mathematical question, I could simply use a calculator. My point is that I actually did use a calculator (and some other methods like Wolfram Alpha and closed source AI models) and I ended up with a completely different result to the OP's math problem than the one you mentioned. That brings me back to my original question: Is it a joke? Some sort of irony or maybe something else entirely that I didn't catch? Honest question. Maybe instead of the over-the-top speculation about trolls that you went into, you could just answer this simple question – how did you arrive at that result? Because that result certainly doesn't seem to be a result to the same math problem posted by OP. Thank you in advance.
That was really funny thanks!
That was really funny thanks!
Hey, I never claimed to be the sharpest tool in the shed, but I'm trying to learn so that I could understand this better so that we could all arrive at the same page here. You seem like you know what's going on. Care to explain? Do I live in a different universe where math works differently? Was that some sort of trick question I didn't catch? Am I missing some important detail that prevents me from understanding it correctly, making me believe that particular result was wrong? Or is it really just someone claiming that the AI model was able to give the correct answer to the OP's math problem, backing it up with the answer from the AI model, but they couldn't be bothered to confirm that by using a calculator before making such claim and then mockingly pointing out that I could always use the calculator to get the same result?
Maybe they are simply ruining the model on the calculator and that's already too many issues from there.
You made my day @MrDevolver