What is BLEU (bilingual evaluation understudy)?

BLEU (bilingual evaluation understudy)

BLEU (bilingual evaluation understudy): A metric scoring machine translation by comparing overlapping word sequences between the machine output and human reference translations.

Related terms

ROUGE (recall-oriented understudy for gisting evaluation)
Dangerous capability evaluation
Evaluation harness
Guardrail evaluation
Holistic Evaluation of Language Models (HELM)
Human evaluation
Model evaluation for dangerous capabilities
Offline evaluation

Browse the full glossary →