{
 "id": "https://redfinchadvisory.com/lexicon/eval",
 "slug": "eval",
 "prefLabel": "Eval",
 "altLabel": [
  "Evaluations",
  "Model evaluation",
  "AI testing"
 ],
 "hiddenLabel": [
  "Vibe check",
  "Vibes-based eval"
 ],
 "definition": "A measurement of how well an AI system does something (tone, helpfulness, accuracy, safety), usually as a score, graded by code, a model or a person.",
 "scopeNote": "An eval doesn't have to be structured or a comparison: a one-off tone check on a single customer-service transcript is an eval. \nNot limited to models; you can eval a harness, a context setup, or other AI infrastructure. Comparison and benchmarking are uses of evals, not the definition. \nEdge case: an eval can have a deterministic answer (did the model solve the math problem?) and still be an eval, because it measures capability.",
 "tier": "core",
 "group": "Evaluation",
 "broader": [],
 "narrower": [
  "https://redfinchadvisory.com/lexicon/benchmark",
  "https://redfinchadvisory.com/lexicon/llm-as-judge",
  "https://redfinchadvisory.com/lexicon/regression-eval",
  "https://redfinchadvisory.com/lexicon/verified-test-set",
  "https://redfinchadvisory.com/lexicon/a-b-test",
  "https://redfinchadvisory.com/lexicon/human-as-judge",
  "https://redfinchadvisory.com/lexicon/production-monitoring",
  "https://redfinchadvisory.com/lexicon/side-by-side-eval"
 ],
 "related": [
  {
   "id": "https://redfinchadvisory.com/lexicon/red-teaming",
   "type": "related"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/observability",
   "type": "related"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/reward-hacking",
   "type": "related"
  }
 ],
 "links": [
  {
   "id": "https://redfinchadvisory.com/lexicon/agent",
   "label": "test",
   "direction": "out"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/verified-test-set",
   "label": "used by",
   "direction": "in"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/bias",
   "label": "measured by",
   "direction": "in"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/evals-engineer",
   "label": "owns",
   "direction": "in"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/probabilistic",
   "label": "is why you need",
   "direction": "in"
  },
  {
   "id": "https://redfinchadvisory.com/lexicon/ai-product-manager",
   "label": "sets the bar with",
   "direction": "in"
  }
 ],
 "modified": "2026-10-21",
 "markdown": "https://redfinchadvisory.com/lexicon/eval.md",
 "inScheme": {
  "id": "https://redfinchadvisory.com/lexicon",
  "version": "1.0",
  "currentAsOf": "2026-10-21"
 }
}