AI response evaluation and comparative LLM output assessment
TextEvaluated AI-generated outputs for quality using criteria such as logical consistency, factual accuracy, readability, and tone alignment. Conducted comparative analysis across multiple large language models to judge instruction-following and response coherence. Used prompt refinement and structured review to improve clarity, structure, and accuracy of generated text for better end-user alignment. • Assessed logical consistency and coherence of model responses • Verified factual accuracy using research and fact-checking methods • Reviewed readability and tone alignment to target audiences • Compared outputs from ChatGPT, Claude AI, and Perplexity AI for relative performance
2024 - Present