Passage Retrieval On Peerqa
평가 지표
MRR
Recall@10
평가 결과
이 벤치마크에서 각 모델의 성능 결과
모델 이름 | MRR | Recall@10 | Paper Title | Repository |
|---|---|---|---|---|
| SPLADEv3 | 0.4536 | 0.6851 | SPLADE-v3: New baselines for SPLADE | |
| Dragon+ | 0.4845 | 0.6817 | How to Train Your DRAGON: Diverse Augmentation Towards Generalizable Dense Retrieval | |
| BM25 | 0.4288 | 0.6388 | - | - |
| Contriever | 0.3624 | 0.5567 | Unsupervised Dense Information Retrieval with Contrastive Learning | |
| MiniLM-L12-v2 | 0.4839 | 0.6709 | MiniLLM: Knowledge Distillation of Large Language Models | |
| GTR-XL | 0.4142 | 0.6122 | Large Dual Encoders Are Generalizable Retrievers | |
| Contriever-MS | 0.4408 | 0.6314 | Unsupervised Dense Information Retrieval with Contrastive Learning | |
| ColBERTv2 | 0.4122 | 0.6371 | ColBERTv2: Effective and Efficient Retrieval via Lightweight Late Interaction |
0 of 8 row(s) selected.