LLM-as-judge — when the judge is another model (step 2/9) · eval-driven ai development

Pairwise asks "A or B better?" → returns A/B/tie. Rubric asks "does this output pass?" → returns pass/fail. Which TWO scenarios above are the strongest fit for pairwise?

⌘↵ runs the editor.read, then continue.

promptdojo_›phase 04 · shipping discipline›ch 21 · eval-driven ai development

lesson 2 of 3 · llm-as-judge — when the judge is another modelstep 2 / 9

Pairwise asks "A or B better?" → returns A/B/tie. Rubric asks "does this output pass?" → returns pass/fail. Which TWO scenarios above are the strongest fit for pairwise?

⌘↵ runs the editor.read, then continue.

LLM-as-judge — when the judge is another model — step 2 of 9

LLM-as-judge — when the judge is another model — step 2 of 9