Recommendation on MovieLens 1M (Accuracy, Recall@3, Recall@5)
51.8AccuracyMixRAGRec
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MixRAGRecBackbone=Mistral-7B2026.05 | 51.8 | 78.5 | 87.9 | |
| MixRAGRecBackbone=LLaMA3-8B2026.05 | 50.4 | 79.8 | 88.2 | |
| K-RagRecBackbone=LLaMA3-8B2026.05 | 45.4 | 65.9 | 70.7 | |
| K-RagRecBackbone=Mistral-7B2026.05 | 43.8 | 72.7 | 83.1 | |
| GPT-4oEvaluation Protocol=Zero-shot Inference2026.05 | 42 | — | — | |
| Rec-r1Backbone=LLaMA3-8B2026.05 | 40.2 | 55.8 | 63.4 | |
| G-retrieverBackbone=LLaMA3-8B2026.05 | 39.5 | 64.8 | 74.3 | |
| Deepseek-r1Evaluation Protocol=Zero-shot Inference2026.05 | 39.3 | — | — | |
| TallRecBackbone=Mistral-7B2026.05 | 39.3 | 66.5 | 76.1 | |
| TallRecBackbone=LLaMA3-8B2026.05 | 39.1 | 66.3 | 77.9 | |
| Gemini-2.0-FlashEvaluation Protocol=Zero-shot Inference2026.05 | 38.9 | — | — | |
| G-retrieverBackbone=Mistral-7B2026.05 | 37.2 | 64.1 | 77.5 | |
| Rec-r1Backbone=Mistral-7B2026.05 | 33.8 | 43.3 | 49 | |
| KAPINGBackbone=Mistral-7B2026.05 | 24.9 | — | — | |
| KG-TextBackbone=LLaMA3-8B2026.05 | 23.4 | — | — | |
| KAPINGBackbone=LLaMA3-8B2026.05 | 23.2 | — | — | |
| KG-TextBackbone=Mistral-7B2026.05 | 22.8 | — | — | |
| LLaMA3-8BBackbone=LLaMA3-8B, Evaluation Protocol=Zero-shot Inference2026.05 | 13 | — | — | |
| Mistral-7BBackbone=Mistral-7B, Evaluation Protocol=Zero-shot Inference2026.05 | 10.5 | — | — |