Novelty Score Prediction on Novelty Assessment Dataset
62AccuracyNovelty Reviewer
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Novelty ReviewerBackbone=Llama-3.1-8B-Instruct2026.01 | 62 | 37.77 | 31.87 | 32.31 | 0.3121 | |
| GPT-OSS-20B2026.01 | 53 | 20.53 | 29.5 | 22.68 | 0.16 | |
| Paper Reviewer2026.01 | 33 | 17.2 | 28.93 | 16.13 | 0.0915 | |
| Llama-3.1-8B-Instruct2026.01 | 15 | 16.21 | 27.97 | 8.94 | 0.0828 | |
| OpenReviewerBackbone=Llama-3.1-8B-Instruct2026.01 | 8 | 19.91 | 20.26 | 8.5 | 0.2412 | |
| Mistral-7B-Instruct-v0.12026.01 | 7 | 15.37 | 25.74 | 3.89 | 0.0746 | |
| SciLlama2026.01 | 6 | 21.21 | 25.31 | 2.94 | -0.0321 | |
| Qwen2.5-14B-Instruct-1M2026.01 | 5 | 26.2 | 25.16 | 2.61 | -0.0321 |