Deep Learning 9 min Cross-Entropy Loss Explained: From Logits to LLM Training What is cross-entropy loss? Derive softmax and negative log-likelihood, work a 3-class example, and see why MSE fails for language model classification. Jun 14, 2026
