Training dataset used for GRPO reinforcement learning fine-tuning of
safoura00/qwen3-terminal-grpo.
The dataset contains 8,000 prompts across 22 Linux terminal task categories,
generated by randomly sampling tasks and formatting them with a system prompt
for command generation.
task_id
string
Task category identifier… See the full description on the dataset page:
https://huggingface.co/datasets/safoura00/qwen3-terminal-tasks.