Minor-embedding on Minor-embedding SAT instances
91.4FJ Success RateLlama-8b SFT BestOf8
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Llama-8b SFT BestOf8Time limit=30s, Number of trials=5, SFT=true, Best-of=82026.04 | 91.4 | 23.4 | 13.58 | 4.06 | 10 | |
| GPT5.2-ReasoningTime limit=30s, Number of trials=52026.04 | 87.76 | 31.2 | 14.33 | 4.86 | 15.1 | |
| GPT5.2Time limit=30s, Number of trials=52026.04 | 87.32 | 8.6 | 20.64 | 4.99 | 18.78 | |
| Llama-8b SFT BestOf1Time limit=30s, Number of trials=5, SFT=true, Best-of=12026.04 | 87.32 | 11.2 | 15.2 | 4.97 | 16.33 | |
| RandomTime limit=30s, Number of trials=52026.04 | 87.16 | — | — | 5.05 | 16.73 |