EverywhereSafety/SEAD-SFT-v1 download history

EverywhereSafety/SEAD-SFT-v1 is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 157 times (157 in the last 7 days), and 157 times in total. It ranks #81,351 among datasets by monthly downloads.

SEAD SFT v1 Supervised fine-tuning data for SEAD: A State-Based Perspective on Attack and Defense in Tool-Using Agents. The dataset trains SAGE, SEAD's state-aware defender, to decide whether a pending tool action should pass or block from the visible interaction history and candidate

Models trained on SEAD-SFT-v1

3 models list it as training data.

Open EverywhereSafety/SEAD-SFT-v1 on Hugging Face

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