deberta-v3-base-prompt-injection-v2 — ONNX

ONNX export of protectai/deberta-v3-base-prompt-injection-v2, a prompt injection detection model fine-tuned on DeBERTa-v3-base to detect prompt injections and jailbreak attempts.

Mirrored for use with inference4j, an inference-only AI library for Java.

Original Source

• Repository: Protect AI
• License: apache-2.0

Usage with inference4j

try (PromptInjectionDetector guard = DebertaPromptInjectionDetector.builder().build()) {
    InjectionCheck check = guard.check(userInput);
}

Model Details

Property	Value
Architecture	DeBERTa-v3-base (model_type: deberta-v2)
Task	Text classification (prompt injection detection)
Classes	0: SAFE, 1: INJECTION
Max sequence length	512
Vocabulary size	128,100
Original framework	PyTorch (Transformers)

License

This model is licensed under the Apache License 2.0. Original model by Protect AI.