Theorem Proving on ProofNet (all)
25.3AccuracyDeepSeek-Prover-V1.5-RL + RMaxTS
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-Prover-V1.5-RL + RMaxTSSample budget=4 x 64002024.08 | 25.3 | |
| DeepSeek-Prover-V1.5-SFT + RMaxTSSample budget=4 x 64002024.08 | 24.8 | |
| DeepSeek-Prover-V1.5-SFTSample budget=4 x 64002024.08 | 22.9 | |
| DeepSeek-Prover-V1.5-RLSample budget=4 x 64002024.08 | 22.6 | |
| InternLM2-StepProverSample budget=1 x 32 x 1002024.08 | 18.1 | |
| ReProver2024.08 | 13.8 | |
| DeepSeek-Prover-V1.5-BaseSample budget=32002024.08 | 13.2 |