Job Description
I have a collection of LLM-generated answers that need a careful native-level review before they are shown to end users. Your job is to read each response in Japanese and judge it on four fronts:
• Natural and correct Japanese
• Accuracy and overall helpfulness of the answer
• Safety, flagging any inappropriate or harmful content
• Cultural content and nuances that may not resonate with Japanese readers
For every item I will provide the original prompt, the AI’s reply, and a simple template (Google Sheet or CSV—your choice) where you can score each of the criteria and leave concise comments or suggested rewrites.
The first batch is 300 short responses; more work will follow if we click. Native fluency is essential and prior experience evaluating machine-generated text is a plus, but clear, actionable feedback matters most.
---
來源:Freelancer.com | 出價次數:3 | 剩餘時間:6 日 | 專案類型:按小時