Benchmarking AIs with a simple riddle
Benchmarking AIs with a simple riddle
Asking models: Sally (a girl) has 3 brothers. Each brother has 2 sisters. How many sisters does Sally have?
Sally (a girl) has 3 brothers. Each brother has 2 sisters. How many sisters does Sally have?
The correct answer is 1 - slightly more compared to how many AIs got it right.
Interestingly though, for ChatGPT if you follow up with "Are you sure about that?" it will correct itself and give the appropriate reasoning.
I wonder if someone told CGPT ;)
ChatGTP got it first try for me
"Sally is one of the sisters mentioned in the statement, so she has one sister. The statement doesn't specify any other sisters, so Sally has one sister."
Same, using Chat GPT 4. It explained the steps without prompting, which is different from the single line answer shown in the post too. I got this…
Let's break this down step by step:
- Sally has 3 brothers.
- Each of those brothers has 2 sisters.
Sally is one of those sisters for each of her 3 brothers. Therefore, the second sister that each brother has would be the same other sister.
This means that Sally has only 1 other sister, making a total of 2 sisters in the family (including Sally herself).
So, Sally has 1 sister.
do you mean GPT*?
good 1. ai.ls got it right. the element/riot group chatgpt 3.5 got it wrong lol. http://0x0.st/Hf36.png
