What is LLM-as-Judge?
Using a model to grade outputs against a rubric so evals can score thousands of cases. Trust it only once its grades agree with human labels on a sample, and recheck whenever the judge model or the rubric changes.
verified on 500 g · 8 graded · 1 rejected
LLM-as-Judge means…
Write it in your own words in the app.
In real work
Before trusting the LLM-as-Judge on 5,000 summaries, the team compared its grades with 150 human labels and rewrote the rubric line where they disagreed most.
In the illustration
Before the line runs, a certified test weight rides across the checkweigher and the display matches its stamped value. Only then does it grade the packs on its own at speed.
Run LLM-as-Judge in your model: grade these 20 summaries against the rubric below, quote the line behind each score, and mark 'unsure' instead of guessing. I will check you against human labels.
Try it now
The fastest way to learn a word is to use it. Open this line in the agent you already work in and watch what it does. Change the tool to the one you use.
Go further
01Practice
Make it stick
One term a day, a short quiz, and sentence practice that checks your own words.
02Verify
Prove you know it
Pass the quiz and keep it in a Skills Wallet you can share with anyone who needs proof.
03Career
Find the work that uses it
See which roles lean on this vocabulary, and what they pay in year one.
Learn the next term tomorrow.
Google gives a definition. Prompthaus pairs each term with how it shows up in standups, PRs, prompts, and design reviews. Then it quizzes you until the word sticks. One term a day on iPhone.
