Mathematical Problem Solving on GSM8K (test)
81.5AccuracyConfidence + AXON
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Confidence + AXONBackbone=LLaDA-1.5, Decoding strategy=Confidence + AXON2026.06 | 81.5 | 22.57 | 72.1 | |
| Confidence + AXONCVRBackbone=LLaDA-1.5, Decoding strategy=Confidence + AXONCVR2026.06 | 81.27 | 23.3 | 70.3 | |
| DAWN + AXONBackbone=LLaDA-1.5, Decoding strategy=DAWN + AXON2026.06 | 81.2 | 33.02 | 49.5 | |
| LocalLeap + AXONCVRBackbone=LLaDA-1.5, Decoding strategy=LocalLeap + AXONCVR2026.06 | 81.12 | 27.64 | 58.1 | |
| DAWNBackbone=LLaDA-1.5, Decoding strategy=DAWN2026.06 | 80.82 | 30.55 | 54.2 | |
| LocalLeap + AXONBackbone=LLaDA-1.5, Decoding strategy=LocalLeap + AXON2026.06 | 80.74 | 26.74 | 60.2 | |
| DAWN + AXONCVRBackbone=LLaDA-1.5, Decoding strategy=DAWN + AXONCVR2026.06 | 80.52 | 34.41 | 47.6 | |
| ConfidenceBackbone=LLaDA-1.5, Decoding strategy=Confidence2026.06 | 80.44 | 22.95 | 74.8 | |
| OriginalBackbone=LLaDA-1.5, Decoding strategy=Original2026.06 | 80.36 | 6.7 | 256 | |
| LocalLeapBackbone=LLaDA-1.5, Decoding strategy=LocalLeap2026.06 | 80.06 | 29.46 | 57.3 | |
| Confidence + AXONBackbone=LLaDA-8B-Instruct, Decoding strategy=Confidence + AXON2026.06 | 79.3 | 23.39 | 74.7 | |
| Confidence + AXONCVRBackbone=LLaDA-8B-Instruct, Decoding strategy=Confidence + AXONCVR2026.06 | 79.15 | 24.16 | 72.2 | |
| ConfidenceBackbone=LLaDA-8B-Instruct, Decoding strategy=Confidence2026.06 | 78.39 | 23.48 | 77.7 | |
| OriginalBackbone=LLaDA-8B-Instruct, Decoding strategy=Original2026.06 | 77.94 | 7.14 | 256 | |
| DAWNBackbone=LLaDA-8B-Instruct, Decoding strategy=DAWN2026.06 | 77.94 | 31.95 | 55.3 | |
| LocalLeap + AXONBackbone=LLaDA-8B-Instruct, Decoding strategy=LocalLeap + AXON2026.06 | 77.48 | 27.97 | 62.1 | |
| DAWN + AXONBackbone=LLaDA-8B-Instruct, Decoding strategy=DAWN + AXON2026.06 | 77.41 | 34.33 | 50.6 | |
| LocalLeapBackbone=LLaDA-8B-Instruct, Decoding strategy=LocalLeap2026.06 | 77.33 | 30.55 | 58.9 | |
| DAWN + AXONCVRBackbone=LLaDA-8B-Instruct, Decoding strategy=DAWN + AXONCVR2026.06 | 76.57 | 35.89 | 48.2 | |
| OriginalBackbone=Dream-v0-Instruct-7B, Decoding strategy=Original2026.06 | 76.35 | 4.85 | 256 | |
| Confidence + AXONBackbone=Dream-v0-Instruct-7B, Decoding strategy=Confidence + AXON2026.06 | 76.12 | 18.53 | 61.6 | |
| LocalLeap + AXONCVRBackbone=LLaDA-8B-Instruct, Decoding strategy=LocalLeap + AXONCVR2026.06 | 75.89 | 28.9 | 59.9 | |
| Confidence + AXONCVRBackbone=Dream-v0-Instruct-7B, Decoding strategy=Confidence + AXONCVR2026.06 | 75.82 | 18.44 | 61.2 | |
| ConfidenceBackbone=Dream-v0-Instruct-7B, Decoding strategy=Confidence2026.06 | 75.2 | 19.23 | 62.3 | |
| LocalLeap + AXONBackbone=Dream-v0-Instruct-7B, Decoding strategy=LocalLeap + AXON2026.06 | 75.06 | 21.97 | 51.5 | |
| LocalLeap + AXONCVRBackbone=Dream-v0-Instruct-7B, Decoding strategy=LocalLeap + AXONCVR2026.06 | 74.91 | 22.04 | 51.1 | |
| DAWN + AXONBackbone=Dream-v0-Instruct-7B, Decoding strategy=DAWN + AXON2026.06 | 74.83 | 22.59 | 51.3 | |
| DAWN + AXONCVRBackbone=Dream-v0-Instruct-7B, Decoding strategy=DAWN + AXONCVR2026.06 | 74.53 | 22.67 | 51.7 | |
| LocalLeapBackbone=Dream-v0-Instruct-7B, Decoding strategy=LocalLeap2026.06 | 73.24 | 22.05 | 54.2 | |
| DAWNBackbone=Dream-v0-Instruct-7B, Decoding strategy=DAWN2026.06 | 73.16 | 22.2 | 52.08 |