core practice · coding problem
Softmax and cross-entropy loss
Logits [2,1,0.1], true class index 0. Compute softmax probabilities and −log p(true).
The mechanism
Softmax turns raw scores (logits) into probabilities that sum to 1. Cross-entropy then measures how wrong those probabilities are versus the true answer. Together they give a clean gradient that backprop can work with.
Expected evidence
p≈[0.659,0.242,0.099], loss≈0.417.
Try this drill in the practice workspace →
Hints when you need them
- Subtract max logit before exp for stability.
- Loss = −log p(correct).