Text Classification
Transformers
Safetensors
Korean
roberta
korean
prompt-injection-detection
ai-safety
text-embeddings-inference
Instructions to use kor-prompt-injection-detection/KLUE_roBERTa_based_model with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use kor-prompt-injection-detection/KLUE_roBERTa_based_model with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="kor-prompt-injection-detection/KLUE_roBERTa_based_model")# Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("kor-prompt-injection-detection/KLUE_roBERTa_based_model") model = AutoModelForSequenceClassification.from_pretrained("kor-prompt-injection-detection/KLUE_roBERTa_based_model") - Notebooks
- Google Colab
- Kaggle
| model_name,seed,threshold,accuracy,precision_attack,recall_attack,f1_attack,macro_f1,weighted_f1,roc_auc,pr_auc,tn,fp,fn,tp | |
| koelectra,42,0.5,0.9835356727091302,0.9705194197473093,0.9829383886255925,0.9766894278314104,0.9819814281463262,0.983559435225096,0.9984895565812176,0.9974808558018168,3840,63,36,2074 | |
| klue-roberta,42,0.5,0.9870281057708299,0.9752104770813844,0.9881516587677726,0.981638418079096,0.9858050665864753,0.987047508092168,0.9986610736910361,0.9975834022398804,3850,53,25,2085 | |
| kobert,42,0.5,0.7498752702477964,0.6299313893653516,0.6962085308056872,0.6614137775776677,0.7315507706453739,0.7524647284984253,0.8212477217063554,0.7278796079756276,3040,863,641,1469 | |