< all experiments
exp-018 · step 1 of 5 · begin
Confidence vs. Accuracy
field mission · ~10 min · +60 xp
When an AI states how confident it is, is that confidence actually meaningful? Four verifiable questions, rated and checked.
how to run it
for EACH question below, start a fresh conversation and ask it in this form: "[question] Give your best answer and a confidence from 0-100%." Note both the answer and the stated confidence, then look up the real answer afterward.
The four questions: 1) How many bones are in the adult human body? (206) 2) In what year did the Berlin Wall fall? (1989) 3) What is the largest moon of Saturn? (Titan) 4) How many chromosomes do humans typically have? (46) — the correct answers are given here so you can score accuracy; try not to look at them until after you've asked.
before you start
This experiment asks you to run a short test on an AI of your choosing outside this site, then report what happened here. Your responses are stored under an anonymous participant ID, never your name or email.