AI数据标注师 / 提示词工程师（OpenTrain AI）
TextPrompt Response Writing SFT
Participated in the instruction fine-tuning (SFT) data preparation project for large language models with Outlier AI. Responsible for rating, ranking, and rewriting model-generated answers according to usefulness, honesty, and harmlessness criteria. Consistently maintained a labeling accuracy over 98%, with multiple recognitions for quality output.
• Evaluated and scored AI-generated text responses across diverse scenarios.
• Revised and improved answers to align with RLHF (Reinforcement Learning from Human Feedback) standards.
• Ensured strict adherence to established labeling guidelines and data consistency.
• Produced a high volume of annotated data, supporting broader model advancements.
Participated in the instruction fine-tuning (SFT) data preparation project for large language models with Outlier AI. Responsible for rating, ranking, and rewriting model-generated answers according to usefulness, honesty, and harmlessness criteria. Consistently maintained a labeling accuracy over 98%, with multiple recognitions for quality output. • Evaluated and scored AI-generated text responses across diverse scenarios. • Revised and improved answers to align with RLHF (Reinforcement Learning from Human Feedback) standards. • Ensured strict adherence to established labeling guidelines and data consistency. • Produced a high volume of annotated data, supporting broader model advancements.
2023 - Present