EverywhereSafety/SEAD-SFT-v1
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 action. Paper · Model weights · GitHub code · Project website · Everywhere Safety Dataset overview The release contains one action-level training example per JSONL row. Each example… See the full description on the dataset page: https://huggingface.co/datasets/EverywhereSafety/SEAD-SFT-v1.
| Author / organization | EverywhereSafety |
|---|---|
| License (as declared) | other |
| Downloads | 213 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 28 September 2026 |
| Last updated at source | 29 September 2026 |
Popularity, last 3 days
213 now. Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- mradermacher/SEAD-SAGE-4B-GGUFModel · used to train
- SEAD: A State-Based Perspective on Attack and Defense in Tool-Using AgentsPaper · cites
- EverywhereSafety/SEAD-SAGE-4BModel · used to train
Related papers
More from EverywhereSafety
- EverywhereSafety/murdoku-labDataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.