Glossary · Term

softmax

← all terms

Definition

Plain language

The step that turns a model's raw scores into probabilities that add up to one, so it can weigh or pick its next word.

As stated in the literature

A function that exponentiates and normalizes a vector of logits into a probability distribution; the normalization step inside attention and in token sampling, and a component whose saturation can throttle gradients at the output head.

Why it matters: It is the step that turns raw scores into usable probabilities inside attention and word selection, and when it saturates it can choke off learning.

For example, it takes a model's raw scores for possible next words and converts them into percentages that add up to a hundred so one can be chosen.

Heard on the show

“The softmax weight: those raw scores get converted into shares of a fixed pie that always sums to one hundred percent.”
Episode 198 — The Model That Knows the Answer and Can't Say It

Mentioned in 3 episodes

  1. 198
    The Model That Knows the Answer and Can't Say It
  2. 179
    How DeepSeek Made One User Faster Without Slowing Down the Crowd
  3. 033
    Echo: The Paper Arguing You Never Needed a KV Cache for Retrieval

Related terms