training task data

[CarlHuangNuc]

Hi Authors,

Where i can find this training dataset ?

"/home/ubuntu/data/webvoyager_persona11000.jsonl"

[CarlHuangNuc]

I can not find another training dataset which using in sft scripts "path/to/webvoyager-claudezeroshot-all85websites"

[CarlHuangNuc]

[YifeiZhou02]

Hi,

Thanks for your interest in our work. Because of legal issues, we cannot release any synthetic data generated by proprietary models like Claude. But you should be able to easily prompt them to generate similar tasks by providing tasks in the example data (https://huggingface.co/datasets/yifeizhou/pae-data/tree/main/tasks). Hope this helps.