Skip to content
InfratoApp
Contribute / chinese-llm-benchmark / LLM-as-judge

Scores output with a model.

chinese-llm-benchmark · LLM-as-judge

Under a minute. No account needed. Everything is checked against the source you give before it appears anywhere.

1 · THE ANSWER

Typed to the attribute — a boolean gets Yes/No, a number gets a number input with its unit, a list gets commas. Never a free-text box where the schema already knows the shape.

2 · WHERE YOU READ IT

The server fetches this before storing anything. If it does not resolve, nothing is saved.

3 · THE SENTENCE THAT SAYS IT

Shown on the product page in quotation marks, exactly as you paste it. This is what lets a reader judge the source instead of trusting us.

YOUR NAME OR HANDLE — OPTIONAL

Shown next to the answer if you give one. Accounts are coming; a sourced answer is checkable either way.