Formal Theorem Proving on Ineq-Comp (test)
66.7Ineq-Comp (Seed)DeepSeek-Prover-V2-7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DeepSeek-Prover-V2-7BBudget=642026.05 | 66.7 | 37.3 | 55.9 | |
| DeepSeek-Prover-V2-7BBudget=322026.05 | 64.8 | 34.8 | 53.7 | |
| DeepSeek-Prover-V2-7B + EnsembleBudget=642026.05 | 64.7 | 45.5 | 70.3 | |
| DeepSeek-Prover-V2-7B + EnsembleBudget=322026.05 | 63.4 | 40.8 | 64.4 | |
| EvolProverBudget=322026.05 | 52.2 | 34 | 65.2 | |
| Goedel-Prover-DPO + EnsembleBudget=642026.05 | 46.6 | 17 | 36.5 | |
| Goedel-Prover-DPOBudget=642026.05 | 44 | 15.3 | 34.8 | |
| Goedel-Prover-DPO + EnsembleBudget=322026.05 | 42.3 | 14.5 | 34.3 | |
| Goedel-Prover-DPOBudget=322026.05 | 40.5 | 12.3 | 30.4 |