Pair-wise Research Idea Comparison on D_pair Hard
63AcchardInnoEval
Evaluation Results
| Method | Links | |
|---|---|---|
| InnoEvalBackbone=DeepSeek-V3.22026.02 | 63 | |
| ScholarEvalBackbone=DeepSeek-V3.22026.02 | 60 | |
| InternAgentBackbone=DeepSeek-V3.22026.02 | 59.5 | |
| GraphEvalBackbone=DeepSeek-V3.22026.02 | 44.5 | |
| ResearchAgentBackbone=DeepSeek-V3.22026.02 | 43 | |
| CoTBackbone=DeepSeek-V3.22026.02 | 40.5 | |
| RAGBackbone=DeepSeek-V3.22026.02 | 39 |