The most basic fact that differentiates the two terms is that the probability distribution must always integrate to 1. This is not true for the likelihood function. If you remember this one fact, then you will know the difference.