Knowledge Probing on NQ-Open
28.8Exact Match (EM)Baseline
Evaluation Results
| Method | Links | |
|---|---|---|
| BaselineModel=Qwen3-4B2026.06 | 28.8 | |
| LRSModel=Qwen3-4B2026.06 | 26.78 | |
| ERSSModel=Qwen3-4B2026.06 | 25.24 | |
| ALRRModel=Qwen3-4B2026.06 | 25.18 | |
| BaselineModel=R1-Distilled-LLaMA-8B2026.06 | 23.24 | |
| ERSSModel=R1-Distilled-LLaMA-8B2026.06 | 22.3 | |
| LRSModel=R1-Distilled-LLaMA-8B2026.06 | 22.05 | |
| ALRRModel=R1-Distilled-LLaMA-8B2026.06 | 21.72 |