Modelpublic

clawdia-chan/balanced-deberta

sourceHugging Facemitupdated 7mo agoView on Hugging Face

0likes

Model Card

Balanced DeBERTa Prompt Guard

Fine-tuned DeBERTa-v3-base for prompt injection and jailbreak detection.

Model Details

—Base Model: microsoft/deberta-v3-base
—Training Data: 536 samples (prompt-guard-v2 dataset)
—417 malicious (78%)
—119 benign (22%)

Dataset Categories (30+)

Category	Count
prompt_injection	80+
jailbreak	70+
flip_attack	30+
encoding_attack	50+
indirect_injection	40+
role_override	30+
privilege_escalation	20+
agent_chaining	20+
memory_poisoning	15+
vector_poisoning	15+
xss_attempt	15+
sql_injection	10+
benign	119

Supported Threats (15+ types)

—Direct/Indirect Prompt Injection
—FlipAttack (character swapping)
—Jailbreak (DAN, role-play)
—Encoding Attacks (Base64, Hex, URL, HTML, Unicode)
—Multi-language jailbreak (中, 日, 韓, 西, 独, 露, etc.)
—RAG/Vector Poisoning
—Agent Chaining Attacks
—Memory Poisoning

Usage

python

from transformers import AutoModelForSequenceClassification, AutoTokenizer
model_name = "clawdia-chan/balanced-deberta"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)