Retrieval-Augmented Generation on TheoremQA
66.3AccuracyOurs
Evaluation Results
| Method | Links | |
|---|---|---|
| OursModel=Qwen3-8B, Retrieval Selection=top-12026.02 | 66.3 | |
| ReFeedLModel=Qwen3-8B, Retrieval Selection=top-12026.02 | 65.1 | |
| InstructRAGModel=Qwen3-8B, Retrieval Selection=top-12026.02 | 64.4 | |
| DPRModel=Qwen3-8B, Retrieval Selection=top-12026.02 | 63.6 | |
| BM25Model=Qwen3-8B, Retrieval Selection=top-12026.02 | 62.8 | |
| ZeroModel=Qwen3-8B, Retrieval Selection=top-12026.02 | 54.6 | |
| OursModel=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 40.8 | |
| ReFeedLModel=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 40 | |
| InstructRAGModel=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 39.2 | |
| DPRModel=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 38.4 | |
| BM25Model=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 37.6 | |
| ZeroModel=Llama-3.1-8B, Retrieval Selection=top-12026.02 | 26.8 |