I wonder how this would compare to a random sample of humans choosing a number between 1 and 100. I assume 42 is frequent due to showing up in the training data a good amount (Answer to "Life, the Universe, and Everything", Jackie Robinson's number and the movie about him), but would that correlate well to how humans would pick a random number?
The "37" is the most common answer for how humans pick a random number.