Evaluation Metric

Quantitative measure of model performance, such as accuracy, F1, BLEU, ROUGE, or perplexity.

Early Stopping

Regularization technique that halts training when validation performance stops improving.

Encoder

Neural network component that maps inputs to latent representations; often bidirectional in Transformers.

Embedding Model

Model that converts inputs to vector embeddings used for search, clustering, or retrieval‑augmented generation.

Embedding

Dense vector representation of text, images, or audio capturing semantic relationships.