Prompt Distillation for Efficient LLM-based Recommendation

Keyphrases

• Distillation 100%
• Immediate Response 20%
• Inference Efficiency 40%
• Inference Time 20%
• Large Language Models 100%
• Long Text 20%
• Modeling Capability 20%
• Modeller 20%
• Multi-step Reasoning 20%
• Noisy Information 20%
• Plaintext 20%
• Recommendation Model 40%
• Recommendation Tasks 20%
• Recommender Systems 20%
• Sequential Recommendation 20%
• Specific Task 20%
• Top-N Recommendation 20%
• Training Effectiveness 40%
• Training Strategy 20%
• User-item 20%

Computer Science

• Experimental Result 20%
• Large Language Model 100%
• Recommender System 20%

Engineering

• Experimental Result 20%
• Large Language Model 100%
• User Model 20%