Encyclopedia Evalica / Evaluation / Eval harness
Eval harness
/ee.val 'hah.rnuhs/The code and infrastructure that runs an eval end-to-end (dataset loading, task execution, scoring, reporting). A harness makes evals repeatable across environments. (noun)
“The eval harness makes the CI run reproducible across branches.”
Related Evaluation terms
- Absolute scoring •
- Agent •
- AI eval •
- Alignment •
- Annotation schema •
- Baseline •
- Baseline experiment •
- Benchmark •
- Calibration •
- CI/CD integration •
- Coherence •
- Confidence interval •
- Eval leakage •
- Experiment •
- Factuality •
- Failure mode •
- Faithfulness •
- Feedback signal •
- Groundedness •
- Hallucination •
- Inter-annotator agreement (IAA) •
- LLM-as-a-judge •
- Loop •
- Model comparison •
- Multimodal •
- Non-determinism •
- Offline evaluation •
- Pairwise evaluation •
- Pass@k •
- Playground •
- Quality gate •
- RAG (retrieval-augmented generation) •
- RAG evaluation •
- Reference-based scoring •
- Reference-free scoring •
- Regression testing •
- Release criteria •
- Remote evaluation •
- Rubric •
- Safety •
- Score distribution •
- Scorer •
- Semantic failure •
- Signal-to-noise ratio •
- Task (eval task) •
- Toxicity score
From the docs
Get started with Evals
Braintrust is the observability platform for agents in production. By actively applying intelligence to agent traces and automatically surfacing critical patterns, Braintrust helps teams at Notion, Stripe, Box, OpenAI, and Cloudflare ship quality agents at scale.
Start building