# LLM-as-judge

Band: established — Established as a technique with well-documented biases — not as a substitute for a deterministic gate.
Group: Verification and authority
Implemented by: https://clickai.dev/skills/advisory-board

Using a model to grade another model's output against a rubric, either as an evaluation or as an inline gate. Now standard and economically viable, with limitations severe enough that any honest description leads with them: position and presentation bias can move accuracy by more than ten points on code evaluation specifically, and it is worst exactly when candidates are close in quality; scores run systematically optimistic. Treat judges as signal and deterministic checks as gates.

## See also

- [Verification gate](https://clickai.dev/legend/verification-gate)
- [Oracle](https://clickai.dev/legend/oracle)
- [Adversarial review](https://clickai.dev/legend/adversarial-review)
- [Echo risk](https://clickai.dev/legend/echo-risk)

---

Full legend: https://clickai.dev/legend.md