Multi-hop Reasoning on HotpotQA (Accuracy, F1)
59.64AccuracyReLAR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ReLARShot Setting=5-shot2026.06 | 59.64 | 76.15 | |
| ReLARShot Setting=0-shot2026.06 | 57.5 | 75.23 | |
| Llama-3-8B-InstructShot Setting=5-shot2026.06 | 56.47 | 71.83 | |
| Med42-Llama3-8BShot Setting=5-shot2026.06 | 54.01 | 69.82 | |
| Qwen2.5-7BShot Setting=5-shot2026.06 | 52.64 | 66.83 | |
| Mistral-7B-InstructShot Setting=5-shot2026.06 | 50.83 | 64.47 | |
| Qwen2.5-Med-7BShot Setting=5-shot2026.06 | 50.18 | 64.15 | |
| Llama-3-8B-InstructShot Setting=0-shot2026.06 | 48.83 | 63.47 | |
| Mistral-7B-v0.3Shot Setting=5-shot2026.06 | 47.82 | 61.35 | |
| MedGemma-4BShot Setting=5-shot2026.06 | 46.91 | 60.05 | |
| Med42-Llama3-8BShot Setting=0-shot2026.06 | 46.23 | 61.59 | |
| Qwen2.5-7BShot Setting=0-shot2026.06 | 44.83 | 58.47 | |
| Gemma-7BShot Setting=5-shot2026.06 | 42.64 | 56.83 | |
| Qwen2.5-Med-7BShot Setting=0-shot2026.06 | 41.81 | 55.13 | |
| Mistral-7B-InstructShot Setting=0-shot2026.06 | 41.47 | 55.28 | |
| LLaMA-2-7BShot Setting=5-shot2026.06 | 39.82 | 52.63 | |
| Med42-Mistral-7BShot Setting=5-shot2026.06 | 39.64 | 53.47 | |
| MedGemma-4BShot Setting=0-shot2026.06 | 39.48 | 53.29 | |
| Mistral-7B-v0.3Shot Setting=0-shot2026.06 | 35.63 | 48.74 | |
| Med42-Mistral-7BShot Setting=0-shot2026.06 | 32.47 | 45.83 | |
| Gemma-7BShot Setting=0-shot2026.06 | 31.82 | 43.47 | |
| Falcon-7BShot Setting=5-shot2026.06 | 30.84 | 41.64 | |
| LLaMA-2-7BShot Setting=0-shot2026.06 | 27.34 | 38.47 | |
| Falcon-7BShot Setting=0-shot2026.06 | 22.47 | 31.83 |