Wiki
Advanced11 min read

Judges and verification

When there is no exact answer, use an LLM-as-judge or a debate between models — carefully. A judge is a heuristic, and it inherits the biases of the models it grades.

Many agent outputs have no exact answer: a summary, a plan, a diagnosis. For these you cannot write a test, so you use a judge — often another model — to score the result against a rubric. It is the only scalable option, and it is a heuristic that can be wrong in systematic ways.

Start here

A judge is a model, so it has the same biases as the model it grades: it favours longer answers, its own style, and confident-sounding text. Use it, but treat its scores as noisy evidence — and anchor it with rubrics, references, and as much mechanical checking as you can.

One model, five benchmarks — same checkpoint, five different stories. Click each for its caveat.

Multiple-choice: format-coachable. Widely cited, easy to game via training data overlap.

Scores are illustrative; the per-benchmark caveats are the lesson. Always ask what a benchmark measures — and whether the model saw its answers during training.

LLM-as-judge

  • Pointwise — score one output against a rubric. Cheap; drifts between runs unless the rubric is tight.
  • Pairwise — show two outputs and ask which is better. More reliable than absolute scores, which is why arenas use it.
  • Known biases — position bias (favouring the first option), length bias (longer looks better), and self-preference (favouring its own family's style).

Debate and verifiers

  • Multi-agent debate — several models answer, critique each other, and converge; the agreement is often more reliable than any single answer.
  • Verifier models — a separate model trained to check a specific property (does this code compile, does this claim follow). Narrow and mechanical beats general and vague.
  • Wherever you can, verify mechanically — a test, a schema, a tool that disagrees. A ground-truth check beats any judge.

Careful

Judges are gameable. Optimising an agent against a judge teaches it to please the judge, not to be right — the reward-hacking story from the LLM domain, applied to evaluation itself. Keep a human-audited sample, measure the judge against human labels, and never let a judge be the only thing standing between a change and production.

Check yourself

Eduspheria wiki · Agentic AI, Evaluation & observability

0 / 4 answered

  1. 1Which judging mode is more reliable than absolute scores, and is used by arenas?
    Multiple choice
  2. 2An LLM judge is free of the biases of the model it grades.
    True / false
  3. 3Which judging bias favours the option shown first?
    Short answer
  4. 4What beats any judge when it is available?
    Multiple choice

Next: seeing what an agent actually did — tracing.