Instructor — Outlier AI
Evaluated AI-generated responses for reasoning quality, coherence, and safety alignment using structured rubric-based guidelines. Rated model outputs for consistency and accuracy while verifying adherence to task instructions. Applied detailed judgment to identify issues that could impact response quality for end users. • Assessed reasoning quality and logical consistency of generated answers • Checked coherence, clarity, and alignment with the prompt • Verified safety alignment based on provided safety guidelines • Ensured ratings matched the expected rubric criteria