AI engineering glossary · Evals guardrails AI security
What is self-preference bias?
Self-preference bias: The tendency of a language model acting as grader to rate text written by itself, or by models of the same family, higher than equally good text from other models. It undermines fairness when a vendor's model judges its own outputs, so teams often use a different model family as judge.
Self-preference bias: an example
A model judges two summaries of equal quality and prefers the one it generated 65 percent of the time, so the team switches to a judge from a different provider.
Related terms
More Evals guardrails AI security terms
- Sensitive information disclosure
- Sleeper agents
- Spotlighting
- STRIDE threat model
- Structured output validation
- SWE-bench
- Synthetic test data
- System prompt extraction
Where you learn it
Self-preference bias is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.