Scientific Proposal Generation on AI Idea Bench and LiveIdeaBench 24 held-out benchmark groups 2025
3.04NoveltyEIG
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| EIG2026.05 | 3.04 | 3.78 | 3.88 | 3.74 | 3.15 | 4.01 | 1.64 | |
| Graph of Thoughts2026.05 | 2.94 | 2.9 | 3.06 | 3 | 2.99 | 2.93 | 4.86 | |
| Self-Refine2026.05 | 2.9 | 2.79 | 3.03 | 3.07 | 3.06 | 2.79 | 5.51 | |
| VirSci2026.05 | 2.9 | 3.14 | 3.71 | 3.28 | 2.99 | 3.5 | 2.92 | |
| AI-Researcher2026.05 | 2.86 | 3.06 | 3.42 | 3.01 | 2.78 | 3.06 | 4.18 | |
| SciPIP2026.05 | 2.71 | 3.1 | 3.5 | 3.08 | 3.08 | 3.31 | 3.15 | |
| Direct2026.05 | 2.18 | 2.72 | 2.76 | 2.71 | 2 | 2.64 | 5.74 |