AI Trainer (described) via Physics expertise, response evaluation, and quality-focused feedback
Performed domain-focused quality checks on science/physics explanations by identifying factual errors, logical inconsistencies, and unclear reasoning in responses. Practiced iterative improvement of explanations through feedback and re-teaching, consistent with the goals of RLHF data creation where responses are scored and refined. Used strong written communication to ensure outputs met clarity, correctness, and instruction-following expectations. • Identified factual and logical issues in explanations • Scored/assessed response quality based on criteria • Provided feedback to improve response clarity and reasoning • Focused on structured, rubric-compliant output writing