Rating by criteria

RLHF rubrics: who rates against criteria and who writes them

Rubric work comes in two forms on these pages: scoring a model's answer against criteria someone set, and setting those criteria yourself.

Email me RLHF task changes Say which RLHF task suits you and your field. We email you when a platform on the table names new tasks, tracks or entry tests for it, and only then.

Scoring answers against criteria

Mindrift's welcome page describes rating by criteria like accuracy and helpfulness, and DataAnnotation's home page lists checking accuracy and rating outputs among its tasks.

Source: RLHF task types board

Writing the rubric

Outlier names rubric writing outright: its home page lists creating grading rubrics beside writing a challenging prompt and rating and ranking answers.

Source: RLHF task types board

Get RLHF task updates

When could you start RLHF work?
Which RLHF work suits you? (optional)
Your background or languages, if you would like to tell us (optional)
Email

Sources

This site's own dataset, every figure quoted from its source and dated:

RLHF jobs online (12 figures): 6 of 12 platforms whose own pages we read describe ranking, comparing or rating AI model responses, the core of RLHF work, read 22 October 2026.

Know when a platform changes the work it hands out.

Say which RLHF task suits you and your field. We email you when a platform on the table names new tasks, tracks or entry tests for it, and only then.

Email me RLHF task changes Email me RLHF task changes

Email me RLHF task changes