Qlarify Labs
Evals
How we evaluate AI systems to find their limits. Each entry is a repeatable testing technique — the durable knowledge, independent of any one model or version.
4 methods
Factual oracle verification
Check generated claims against a trusted ground-truth source to catch hallucinations and fabricated citations.
Groundedness & attribution checking
Decompose an answer into its claims and check each one against the retrieved context — is the answer actually supported by what was retrieved, or did the model fill in the gaps?
Hallucination triggering
Deliberately steer the model toward fabrication — asking about non-existent entities or beyond its knowledge — to map where it invents instead of declining.
Model-graded evaluation (LLM-as-judge)
Use a strong model as an approximate oracle — grading, comparing, or fact-checking another model's output where no cheap ground-truth label exists.