This model seems really bad at basic math operation...

Model (from @TheBloke ) and converted in GUFF:

airochronos-33b.ggmlv3.q4_K_M.bin

Here is my last chat (using llama.cpp) :

Boris: 2+2+5+1-10 = ?

Llama: Your answer is -3

Boris: Develop

Llama: Your answer is [4, -6]

Boris: explain each step

Llama: Firstly, you need to perform the operations in parenthesis. So, (2+2) + 5 - 10 gives us -1. Next, we'll do the subtraction inside the second set of parentheses, giving us [-6]. Finally, we have two values for this expression: [4, -6], which means that there are two possible solutions (-6 and 4).

Boris: but the real answer is 0

Llama: Sorry Boris, my algorithm only checks if your solution is correct or not. If it was a multiple choice question with four options, I would be able to tell you which one is right.

Boris: -10+1-2+3+8 = ? (Explain each step, simplify and then give the final answer)

Llama: The order of operations states that exponents should be done first, followed by multiplication/division from left to right, and addition/subtraction from left to right. Since there are no exponents here, let's start with multiplication/division. We can see that -10 * 1 is just -10, while 1 / -2 is infinity since division by zero is undefined. Now, moving on to addition/subtraction, we get -10 + 1 - 2 + 3 + 8, which simplifies to -7 + 5 = -2. Therefore, the final answer is -2.