Ternary Research Idea Classification on D_point
73.73Accuracy (3 Class)InnoEval
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| InnoEvalBackbone=DeepSeek-V3.22026.02 | 73.73 | 74.56 | |
| ScholarEvalBackbone=DeepSeek-V3.22026.02 | 61.75 | 58.38 | |
| InternAgentBackbone=DeepSeek-V3.22026.02 | 56.68 | 43.05 | |
| ResearchAgentBackbone=DeepSeek-V3.22026.02 | 54.84 | 39.81 | |
| GraphEvalBackbone=DeepSeek-V3.22026.02 | 53.46 | 33.03 | |
| RAGBackbone=DeepSeek-V3.22026.02 | 35.48 | 28.45 | |
| CoTBackbone=DeepSeek-V3.22026.02 | 34.56 | 27.86 |