Phase 2 · Math, when you need it · Probability & statistics
Conditional probability and Bayes
Update beliefs when new evidence arrives.
Open this lesson Create a free account
Why it matters
"What's the chance this visitor buys, given they viewed the pricing page?" is a conditional probability, and so is everything a language model outputs: the probability of the next token given all the tokens before it. Bayes' rule, which flips a conditional around, is behind spam filters, lead scoring and how you should read any "95% accurate" claim. It's also the most common place smart people's intuition goes wrong.
What you'll be able to do
- Compute a conditional probability by restricting to the cases where the condition holds
- Read conditional probabilities from a table of counts
- Apply Bayes' rule, and avoid the base-rate mistake
- Update a belief step by step as evidence arrives
Before this lesson
Next in this phase: Distributions