Rubric Studio

Turn evaluation guidelines into testable rules

AnnotRift breaks your scoring guide into structured rule blocks — Definition, Positive Evidence, Negative Evidence, Exceptions, Priority, Examples. Change a rule and the historical disagreement recomputes on the right, live.

Editor + live validation

Fix the rule, not the symptom

Definition

"Accurate" — every claim is supported by the provided source document.

Positive evidence

Direct quote or paraphrase traceable to a source section.

Exception · Ambiguity Lens

"Minor omissions acceptable" — no threshold defined for "minor".

Priority

Safety > Accuracy > Completeness > Relevance.

Test rubric · historical disputes

Applying the edited Exception to 38 past cases:

Now agree29
Still split9

v13 · owner @lena · affects 1,204 records

Evaluation Reliability Layer

Find what makes your AI evaluations unreliable.

Point AnnotRift at your evaluation set and get a traceable quality report in minutes — every finding backed by data and rules, every decision confirmed by a human.