Rollouts from 4B and 14B models logged during their RL post-training: https://huggingface.co/datasets/OpenHands/CodeScout_Training_Rollouts Rollouts from CodeScout 14B used to warm-start the 1.7B model with RFT: https://huggingface.co/datasets/adityasoni17/CodeScout14B_RFT_SWE_Smith
Rollouts from 4B and 14B models logged during their RL post-training: https://huggingface.co/datasets/OpenHands/CodeScout_Training_Rollouts
Rollouts from CodeScout 14B used to warm-start the 1.7B model with RFT: https://huggingface.co/datasets/adityasoni17/CodeScout14B_RFT_SWE_Smith