RLHF annotator and LLM logic evaluator
Served as a specialist in RLHF annotation and model evaluation for large language models. Assessed business logic, reviewed bilingual content, and evaluated model output for accuracy, reasoning, and factual correctness. Identified logical inconsistencies in commercial scenarios and contributed to knowledge extraction processes. • Involved in complex text data evaluation and annotation. • Performed RLHF (Reinforcement Learning from Human Feedback) tasks specific to business and finance domains. • Conducted bilingual review in Chinese and English for LLM output. • Supported quality calibration and assessment activities for AI models.