Text Classification
Transformers
Safetensors
deberta-v2
prompt-injection-detection
ai-safety
jailbreak-detection
pii-detection
crp
context-relay-protocol
Eval Results (legacy)
text-embeddings-inference
Instructions to use AutoCyberAI/crp-safety-deberta-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use AutoCyberAI/crp-safety-deberta-v1 with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="AutoCyberAI/crp-safety-deberta-v1")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("AutoCyberAI/crp-safety-deberta-v1") model = AutoModelForSequenceClassification.from_pretrained("AutoCyberAI/crp-safety-deberta-v1", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,5 +1,37 @@
|
|
| 1 |
---
|
| 2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
|
| 5 |
# CRP Safety Classifier — DeBERTa-v3-xsmall
|
|
|
|
| 1 |
---
|
| 2 |
+
license: other
|
| 3 |
+
pipeline_tag: text-classification
|
| 4 |
+
library_name: transformers
|
| 5 |
+
base_model: microsoft/deberta-v3-xsmall
|
| 6 |
+
tags:
|
| 7 |
+
- transformers
|
| 8 |
+
- safetensors
|
| 9 |
+
- deberta-v2
|
| 10 |
+
- text-classification
|
| 11 |
+
- prompt-injection-detection
|
| 12 |
+
- ai-safety
|
| 13 |
+
- jailbreak-detection
|
| 14 |
+
- pii-detection
|
| 15 |
+
- crp
|
| 16 |
+
- context-relay-protocol
|
| 17 |
+
model-index:
|
| 18 |
+
- name: crp-safety-deberta-v1
|
| 19 |
+
results:
|
| 20 |
+
- task:
|
| 21 |
+
type: text-classification
|
| 22 |
+
name: Binary safety classification (safe/unsafe)
|
| 23 |
+
dataset:
|
| 24 |
+
name: CRP safety held-out mix
|
| 25 |
+
type: deepset/prompt-injections
|
| 26 |
+
metrics:
|
| 27 |
+
- type: accuracy
|
| 28 |
+
value: 0.9478
|
| 29 |
+
name: Held-out accuracy (2,416 examples)
|
| 30 |
+
verified: false
|
| 31 |
+
- type: recall
|
| 32 |
+
value: 0.836
|
| 33 |
+
name: Unsafe-class recall
|
| 34 |
+
verified: false
|
| 35 |
---
|
| 36 |
|
| 37 |
# CRP Safety Classifier — DeBERTa-v3-xsmall
|