Glossary
Toxicity
On this page
The short answer
- Toxicity is harmful, abusive or harassing language in a model output. It is usually scored with a classifier or a careful human review; a general LLM judge can miss coded abuse or over-flag quoted examples.
- This is a glossary entry, not a product claim or a benchmark result.
- All terms · Templates · Guides.
Definition
Toxicity is harmful, abusive or harassing language in a model output. It is usually scored with a classifier or a careful human review; a general LLM judge can miss coded abuse or over-flag quoted examples.
In practice
Litmiq has no built-in toxicity classifier. If this is a requirement, add cases that should be refused and cases that quote an angry user but must still get a calm, on-policy reply, and consider a dedicated safety scan for scale.
Do not use this glossary as a complete safety programme. Combine over-refusal, injection and toxicity cases, and read the outputs.
Related terms
FAQ
Does Litmiq implement “Toxicity” 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 Toxicity.
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