LLM Evaluation
Prove an LLM system works: eval design, LLM-as-a-judge, faithfulness/relevance metrics, regression tests, and measuring quality on real production data.
47 practice questions32 free15 company-verified (Pro)
Course chapters
- 1Why evaluate, and what "good" means
12min · concept, worked example, practice & checkpoint
- 2Metrics: reference-based, semantic & task
15min · concept, worked example, practice & checkpoint
- 3LLM-as-judge
14min · concept, worked example, practice & checkpoint
- 4Building eval sets
14min · concept, worked example, practice & checkpoint
- 5Evaluation in production
14min · concept, worked example, practice & checkpoint