Glossary
Semantic similarity
On this page
The short answer
- Semantic similarity scores how close an output is in meaning to a reference, often with embeddings or a judge, rather than exact string match. It can accept paraphrases, and it can also reward fluent text that is still wrong on a critical fact.
- This is a glossary entry, not a product claim or a benchmark result.
- All terms · Templates · Guides.
Definition
Semantic similarity scores how close an output is in meaning to a reference, often with embeddings or a judge, rather than exact string match. It can accept paraphrases, and it can also reward fluent text that is still wrong on a critical fact.
In practice
If a single word (a price, a date, a refusal) is what matters, use a deterministic check or expected behavior, not a similarity number alone.
Litmiq does not compute embedding similarity. You can still include a reference in expected behavior and ask the judge to compare.
Related terms
FAQ
Does Litmiq implement “Semantic similarity” as a built-in metric?
Only where the definition itself says so. Many terms are industry vocabulary. Litmiq is a single-turn, BYOK prompt tester with an LLM judge, not a full evaluation platform.
Where is the rest of the glossary?
The hub at /glossary lists every term. This page is the standalone entry for Semantic similarity.
See the term in a suite
Many glossary entries have a matching template you can run in the browser.
- Free, no signup
- Runs in your browser
- Your OpenAI key, sent only to OpenAI