Mathematical Reasoning on MATH 128
31.41AccuracyNI Sampling
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| NI SamplingModel=LLaDA 1.52026.04 | 31.41 | 39.65 | 87.1 | |
| FullModel=LLaDA 1.52026.04 | 31.03 | 128 | 27.8 | |
| ThresholdModel=LLaDA 1.52026.04 | 30.93 | 58.71 | 60.9 | |
| NI SamplingModel=LLaDA-8B-Instruct2026.04 | 30.07 | 39.01 | 87.2 | |
| ThresholdModel=LLaDA-8B-Instruct2026.04 | 29.69 | 60.43 | 57.9 | |
| FullModel=LLaDA-8B-Instruct2026.04 | 29.64 | 128 | 28.5 |