When using my math app it is responding to my answer saying they are wrong when they are right and right when they are wrong. How can I get more accurate responses from my bot.
LLMs are generally not great at hard math. What type of math are you doing exactly?
Basic to pre algebra. So for 3rd graders to 8th graders.
It should be able to do math like this with a fair level of accuracy. But it’s not infallible. These are, at the end of the day, language models and not math models. What exactly is it getting it wrong?
Sometimes it won’t read the answer correctly so it will say it is right when it might be wrong and sometimes when it is right it will act as if it is wrong.
This is the type of responses it is giving.