Code Generation on MBPP (Pass@1, Execution Time)
44.73Pass@1Saber
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SaberSampling Strategy=Efficient DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 44.73 | 0.3302 | 110.96 | 1 | |
| Fast-dLLMSampling Strategy=Efficient DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 44.03 | 0.3044 | 256 | 2 | |
| WINOSampling Strategy=Efficient DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 43.09 | 0.3138 | 88.49 | 1 | |
| ConfidenceSampling Strategy=Standard DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 42.86 | 0.3138 | 256 | 3 | |
| EntropySampling Strategy=Standard DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 42.15 | 0.3114 | 256 | 2 | |
| Confidence (p=2)Sampling Strategy=Efficient DLM Sampling, p parameter=2, Backbone=LLaDA-8B-Instruct2025.10 | 40.75 | 0.2857 | 128 | 1 | |
| SAR (p=2)Sampling Strategy=Efficient DLM Sampling, p parameter=2, Backbone=LLaDA-8B-Instruct2025.10 | 40.05 | 0.2786 | 128 | 1 | |
| Fast-dLLM (+parallel)Sampling Strategy=Efficient DLM Sampling, Parallelism=true, Backbone=LLaDA-8B-Instruct2025.10 | 39.34 | 0.2763 | 73.13 | 43 | |
| ReMDMSampling Strategy=Efficient DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 31.62 | 0.2248 | 128 | 1 | |
| RandomSampling Strategy=Standard DLM Sampling, Backbone=LLaDA-8B-Instruct2025.10 | 22.95 | 0.1826 | 256 | 2 |