This platform is designed for human evaluation of LLM-generated Korean SAT reading comprehension passages. It provides a web-based interface where evaluators can assess the quality and appropriateness of AI-generated passages and questions for the Korean College Scholastic Ability Test (CSAT).
The platform serves as a human evaluation tool for assessing LLM-generated Korean SAT reading comprehension passages. Evaluators can:
- Review AI-generated passages and questions
- Evaluate the quality and difficulty level of passages
- Assess the relevance and appropriateness of questions
- Provide feedback on the overall quality of the content
- Real-time loading of evaluation data from Google Sheets
- Structured evaluation forms for passage and question assessment
- Firebase-based storage of evaluation results
- Real-time feedback collection and analysis
- User-friendly interface for evaluators
- Python
- Streamlit (Web Interface)
- Google Sheets API
- Firebase Admin SDK
- Pandas (Data Processing)