Query-Focused Submodular Demonstration Selection for In-Context Learning in Large Language Models

Computer Science

• Large Language Model 100%
• Interactive Tool 50%
• Performance Improvement 50%
• Performance Model 50%
• Random Selection 50%
• Leaning Parameter 50%
• Predictive Performance 50%
• Mutual Information 50%
• Sentiment Classification 50%
• Parameter Size 50%
• Zero-Shot Learning 50%