Smith-Waterman Alignment Score Prediction on AIME combined 2020-2025
0.368R2GPT-OSS-120B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| GPT-OSS-120BFeature Set=Graph + TDA2025.10 | 0.368 | 0.236 | 12.5 | 8.3 | |
| GPT-OSS-20BFeature Set=Graph + TDA2025.10 | 0.327 | 0.186 | 10.5 | 2.2 | |
| GPT-OSS-120BFeature Set=TDA2025.10 | 0.327 | 0.218 | — | — | |
| Qwen3-8BFeature Set=Graph + TDA2025.10 | 0.312 | 0.167 | 14.3 | 8.4 | |
| DeepSeek-r1-70BFeature Set=Graph + TDA2025.10 | 0.3 | 0.153 | 20.5 | 20.5 | |
| GPT-OSS-20BFeature Set=TDA2025.10 | 0.296 | 0.182 | — | — | |
| Qwen3-8BFeature Set=TDA2025.10 | 0.273 | 0.154 | — | — | |
| DeepSeek-r1-70BFeature Set=TDA2025.10 | 0.249 | 0.127 | — | — | |
| Qwen3-32BFeature Set=Graph + TDA2025.10 | 0.233 | 0.073 | 28.7 | 52.1 | |
| DeepSeek-r1-7BFeature Set=Graph + TDA2025.10 | 0.226 | 0.063 | 7.6 | -23.2 | |
| DeepSeek-r1-32BFeature Set=Graph + TDA2025.10 | 0.226 | 0.064 | 18.9 | 8.5 | |
| DeepSeek-r1-7BFeature Set=TDA2025.10 | 0.21 | 0.082 | — | — | |
| DeepSeek-r1-32BFeature Set=TDA2025.10 | 0.19 | 0.059 | — | — | |
| Qwen3-32BFeature Set=TDA2025.10 | 0.181 | 0.048 | — | — | |
| Qwen3-235BFeature Set=Graph + TDA2025.10 | 0.167 | -0.001 | 2.5 | -103.7 | |
| Qwen3-235BFeature Set=TDA2025.10 | 0.163 | 0.027 | — | — | |
| GPT-OSS-120BFeature Set=Graph2025.10 | 0.101 | 0.07 | — | — | |
| Qwen3-32BFeature Set=Graph2025.10 | 0.088 | 0.056 | — | — | |
| GPT-OSS-20BFeature Set=Graph2025.10 | 0.081 | 0.049 | — | — | |
| DeepSeek-r1-70BFeature Set=Graph2025.10 | 0.058 | 0.025 | — | — | |
| DeepSeek-r1-32BFeature Set=Graph2025.10 | 0.057 | 0.024 | — | — | |
| Qwen3-8BFeature Set=Graph2025.10 | 0.054 | 0.021 | — | — | |
| DeepSeek-r1-7BFeature Set=Graph2025.10 | 0.047 | 0.014 | — | — | |
| Qwen3-235BFeature Set=Graph2025.10 | 0.024 | -0.004 | — | — |