Short answer
What is RLHF work?
RLHF work usually means helping improve model behavior by comparing, rating, or rewriting outputs according to a rubric. Applicants may need strong writing, domain judgment, or technical reasoning.
Context
RLHF stands for reinforcement learning from human feedback. In public listings, the label can cover several task types, so the current role page matters.
Current general AI evaluation examples
These current roles are examples of the kind of AI evaluation work described above. Read each listing because the actual tasks and requirements differ.
Mercor · general
Google Workspace & Business Profile Owners
$60 / hour
Micro1 · general
Data-Video Generalist (US-based)
$13 - $15 / hour
Mercor · general
Payment-posting & Reconciliation Manager
$85 / hour