Question Answering on WebQuestions first 1000 samples (dev)
30.6EMPathfinder
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| PathfinderBase Model=Gemma-2-9B-Instruct, Distribution Setting=out-of-distribution2025.12 | 30.6 | 34.1 | 44.08 | |
| DeepRAG (Imi-retrieve only)Base Model=Qwen-2.5-7B-Instruct, Distribution Setting=out-of-distribution2025.12 | 29.3 | 33.8 | 42.49 | |
| DeepRAG (Imi-retrieve only)Base Model=Gemma-2-9B-Instruct, Distribution Setting=out-of-distribution2025.12 | 29.3 | 34.2 | 41.92 | |
| PathfinderBase Model=Qwen-2.5-7B-Instruct, Distribution Setting=out-of-distribution2025.12 | 28.5 | 38.1 | 42.47 | |
| DeepRAG (Imi)Base Model=Qwen-2.5-7B-Instruct, Distribution Setting=out-of-distribution2025.12 | 25.3 | 28.4 | 38.01 | |
| DeepRAG (Imi)Base Model=Gemma-2-9B-Instruct, Distribution Setting=out-of-distribution2025.12 | 23.6 | 27.7 | 36.18 |