Belief Under Evidence

Interactive probability distribution over sixteen hypotheses. In Bayesian Learning mode, each click adds evidence and the distribution sharpens from uniform ignorance toward peaked knowledge, entropy falling on average as evidence accumulates. In Dogma mode, one click collapses it onto a fixed hypothesis regardless of the evidence.

In Bayesian Learning, every click is an observation and belief sharpens toward where the evidence points, entropy falling a little at a time. In Dogma, a single click collapses belief onto the same fixed hypothesis (H₁₂) no matter where you clicked, dropping entropy to 0.120 bits in one step.