Ranking on MovieLens 1M
0.615NDCG@5FlexRec
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FlexRecBase Model=Qwen2.5-3B-Instruct, Protocol=Post-Trained (Sparse)2026.03 | 0.615 | 0.665 | 0.844 | 0.235 | 0.368 | |
| TALLRecBase Model=Qwen2.5-3B-Instruct, Protocol=Post-Trained (Sparse)2026.03 | 0.569 | 0.609 | 0.821 | 0.184 | 0.33 | |
| Rec-R1Base Model=Qwen2.5-3B-Instruct, Protocol=Post-Trained (Sparse)2026.03 | 0.554 | 0.6 | 0.817 | 0.17 | 0.325 | |
| GPT-4oProtocol=Zero-Shot2026.03 | 0.508 | 0.564 | 0.8 | 0.139 | 0.244 | |
| BERT4RecCategory=Traditional Reranker2026.03 | 0.502 | 0.56 | 0.796 | 0.128 | 0.231 | |
| STARCategory=Traditional Reranker2026.03 | 0.5 | 0.562 | 0.797 | 0.129 | 0.234 | |
| Qwen2.5-3B-InstructProtocol=Zero-Shot2026.03 | 0.497 | 0.553 | 0.794 | 0.131 | 0.245 | |
| Rank-GRPOBase Model=Qwen2.5-3B-Instruct, Protocol=Post-Trained (Sparse)2026.03 | 0.492 | 0.549 | 0.792 | 0.131 | 0.245 |