Dataset · deepset
Prompt injection dataset for detection
deepset/prompt-injections
Small text dataset of prompt injections; card has minimal description.
What's inside
Text dataset with fewer than 1K examples. The card indicates only the name deberta-v3-base-injection-dataset and a More Information needed note; there are no details on collection, language, or annotation.
What you can do with it
- Train a prompt injection classifier
- Check LLM application protection against injections
- Test the DeBERTa model on injection detection
- Size
- up to 1K rows
- Downloads this month
- 15.6K
- Likes
- 202
Models trained on it
- patronus-studio/wolf-defender-prompt-injectionText classification 10K
- testsavantai/prompt-injection-defender-large-v0Text classification 5K
- patronus-studio/wolf-defender-prompt-injection-smallText classification 3.8K
- deepset/deberta-v3-base-injectionText classification 2.2K
- codeintegrity-ai/promptguardText classification 1.4K
- fmops/distilbert-prompt-injectionText classification 1.1K
- mradermacher/Spellguard-1.0-8B-GGUF 934
- mradermacher/monitor-qwen35-08b-GGUFText classification 766
How to load it
from datasets import load_dataset
ds = load_dataset("deepset/prompt-injections")