LLM Response Evaluation & Annotation Practice | Independent Portfolio Project
Evaluated AI-generated answers for instruction following, factual accuracy, clarity, tone, and completeness using detailed rubric criteria. Provided concise, written rationales explaining choices, missing requirements, hallucination risks, and improvement opportunities. Assessed technical prompts across React, JavaScript, APIs, data analysis, and general reasoning for ambiguity and edge cases. • Conducted systematic review of LLM and generative AI model outputs • Focused on consistency and objectivity in annotation • Ensured feedback is actionable for future AI improvement • Materials included both technical and general domain content.