hassanjbara/LONG-DPO download history

hassanjbara/LONG-DPO is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 20 times (2 in the last 7 days), and 1,390 times in total. It ranks #374,093 among datasets by monthly downloads.

This dataset is a DPO version of the hassanjbara/LONG dataset. The preference model used for choosing and rejecting responses is Hello-SimpleAI/chatgpt-detector-roberta, with the idea being creating a dataset for training a model to "beat" that detector. Full script for generating the dataset inclu

Models trained on LONG-DPO

2 models list it as training data.

• hassanjbara/Phi-3-mini-4k-natural – downloads in 30 days
• hassanjbara/Meta-Llama-3.1-8B-natural – downloads in 30 days

Open hassanjbara/LONG-DPO on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.