AI Trainer – RLHF Projects
As an AI Trainer, I participated in reinforcement learning from human feedback (RLHF) projects aimed at improving large language models. My work involved response ranking, evaluation, and high-accuracy judgment of chatbot outputs following detailed rubrics. I provided linguistic and guideline compliance insights for both English and Chinese text datasets. • Evaluated and rated language model responses for quality and relevance. • Followed strict rubrics to ensure fairness and consistency in rankings. • Contributed to multilingual model evaluations for English and Chinese tasks. • Utilized tools like Scale AI and CVAT for annotation and evaluation.