This is the source code used to conduct the experiments described in the blog post Cut the bullshit: Are LLMs able to quantify (un)certainty?