Question Answering on en multifield
44.29F1 ScoreFullKV
Evaluation results
| Method | Links | |
|---|---|---|
| FullKVModel=Llama-3.1-8B-Instruct2026.01 | 44.29 | |
| RAG-RerankModel=Llama-3.1-8B-Instruct2026.01 | 43.55 | |
| HyDE-RerankModel=Llama-3.1-8B-Instruct2026.01 | 43.55 | |
| S3-HybridModel=Llama-3.1-8B-Instruct2026.01 | 43.54 | |
| BM25Model=Llama-3.1-8B-Instruct2026.01 | 43.44 | |
| S3-PureModel=Llama-3.1-8B-Instruct2026.01 | 42.1 | |
| S3-HybridModel=Mistral-7B-Instruct-v0.32026.01 | 38.87 | |
| BM25Model=Mistral-7B-Instruct-v0.32026.01 | 38.87 | |
| HyDE-RerankModel=Mistral-7B-Instruct-v0.32026.01 | 38.73 | |
| RAG-RerankModel=Mistral-7B-Instruct-v0.32026.01 | 38.62 | |
| HyDE-RerankModel=Qwen2-7B-Instruct2026.01 | 38.42 | |
| RAG-RerankModel=Qwen2-7B-Instruct2026.01 | 38.33 | |
| FullKVModel=Mistral-7B-Instruct-v0.32026.01 | 38.24 | |
| BM25Model=Qwen2-7B-Instruct2026.01 | 38 | |
| FullKVModel=Qwen2-7B-Instruct2026.01 | 37 | |
| S3-HybridModel=Qwen2-7B-Instruct2026.01 | 36.75 | |
| S3-PureModel=Mistral-7B-Instruct-v0.32026.01 | 35.52 | |
| S3-PureModel=Qwen2-7B-Instruct2026.01 | 34.87 | |
| RAG-DenseModel=Llama-3.1-8B-Instruct2026.01 | 32.67 | |
| RAG-DenseModel=Mistral-7B-Instruct-v0.32026.01 | 28.66 | |
| RAG-DenseModel=Qwen2-7B-Instruct2026.01 | 27.99 |