Question Answering on HotpotQA (Accuracy)
33.91AccuracyRankCoT
Evaluation Results
| Method | Links | |
|---|---|---|
| RankCoTBackbone=Qwen2.5-14B-Instruct2025.02 | 33.91 | |
| RankCoTBackbone=MiniCPM3-4B2025.02 | 33.13 | |
| RankCoTBackbone=Llama3-8B-Instruct2025.02 | 32.21 | |
| RerankBackbone=Llama3-8B-Instruct2025.02 | 30.3 | |
| No RefinementBackbone=Qwen2.5-14B-Instruct2025.02 | 29.7 | |
| No RefinementBackbone=Llama3-8B-Instruct2025.02 | 29.43 | |
| SummaryBackbone=Llama3-8B-Instruct2025.02 | 28.05 | |
| CoTBackbone=Llama3-8B-Instruct2025.02 | 26.36 | |
| No RefinementBackbone=MiniCPM3-4B2025.02 | 24.93 |