Comment by mohamedkoubaa

Do we know if it in the training data or if the humans giving rewards in the labs have such awful taste in prose?