HF RL Explorer

lgy0404/mobileforge-training-data

mobileforge-training-data: RL dataset on Hugging Face. This dataset contains the hint-contextualized step-level GRPO training data used by MobileForge. The samples are produced from MobileGym rollouts, hierarchical critic feedback, and HiFPO filtering. The Hugging Face Dataset Viewer is disabled…

lgy0404/mobileforge-training-data on the Hugging Face Hub