Multi-document Retrieval on HotpotQA (val)
70.96F1 ScoreGritLM
Evaluation Results
| Method | Links | |
|---|---|---|
| GritLMModel Category=Dense Models, top_k=32026.02 | 70.96 | |
| AttentionRetrieverModel Category=AttentionRetriever (Ours), Backbone=LLaMA-3.2, Model size=3B, top_k=32026.02 | 70.9 | |
| AttentionRetrieverModel Category=AttentionRetriever (Ours), Backbone=Qwen-2.5, Model size=3B, top_k=32026.02 | 70.37 | |
| Qwen3Model Category=Dense Models, top_k=32026.02 | 67.21 | |
| GTE-Qwen2Model Category=Dense Models, top_k=32026.02 | 67.07 | |
| GTRModel Category=Dense Models, top_k=32026.02 | 61.09 | |
| SPScannerModel Category=Autoregressive Models, top_k=32026.02 | 60.52 | |
| BM25Model Category=Sparse Models, top_k=32026.02 | 56.95 | |
| ANCEModel Category=Dense Models, top_k=32026.02 | 55.47 | |
| DPRModel Category=Dense Models, top_k=32026.02 | 50.2 | |
| CDEModel Category=Dense Models, top_k=32026.02 | 39.32 |