Response Generation on University Admission Inquiries (test)
92.7Fact RecallFine-Tuned with RAG
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Fine-Tuned with RAGRetrieval-Augmented=Yes, Fine-tuned=Yes2025.12 | 92.7 | 88.3 | 8.9 | |
| RAG ModelRetrieval-Augmented=Yes, Fine-tuned=No2025.12 | 75.1 | 91.4 | 7.5 | |
| Fine-Tuned (No RAG)Retrieval-Augmented=No, Fine-tuned=Yes2025.12 | 72.7 | 48.3 | 7.9 | |
| Baseline GPTType=Baseline2025.12 | 22.3 | 25.6 | 3.2 |