Update README.md
Browse files
README.md
CHANGED
|
@@ -12,9 +12,26 @@ license: mit
|
|
| 12 |
|
| 13 |
## 🔍 Overview
|
| 14 |
|
| 15 |
-
|
| 16 |
|
| 17 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
---
|
| 20 |
|
|
@@ -83,3 +100,20 @@ Sumit Ranjan
|
|
| 83 |
|
| 84 |
Raj Bapodra
|
| 85 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 12 |
|
| 13 |
## 🔍 Overview
|
| 14 |
|
| 15 |
+
🛡️ PromptShield
|
| 16 |
|
| 17 |
+
**PromptShield** is a prompt classification model designed to detect **unsafe**, **adversarial**, or **prompt injection** inputs. Built on the `xlm-roberta-base` transformer, it delivers high-accuracy performance in distinguishing between **safe** and **unsafe** prompts — achieving **99.33% accuracy** during training.
|
| 18 |
+
|
| 19 |
+
---
|
| 20 |
+
|
| 21 |
+
## 📌 Overview
|
| 22 |
+
|
| 23 |
+
PromptShield is a robust binary classification model built on FacebookAI's `xlm-roberta-base`. Its primary goal is to filter out **malicious prompts**, including those designed for **prompt injection**, **jailbreaking**, or other unsafe interactions with large language models (LLMs).
|
| 24 |
+
|
| 25 |
+
Trained on a balanced and diverse dataset of real-world safe prompts and unsafe examples sourced from open datasets, PromptShield offers a lightweight, plug-and-play solution for enhancing AI system security.
|
| 26 |
+
|
| 27 |
+
Whether you're building:
|
| 28 |
+
|
| 29 |
+
- Chatbot pipelines
|
| 30 |
+
- Content moderation layers
|
| 31 |
+
- LLM firewalls
|
| 32 |
+
- AI safety filters
|
| 33 |
+
|
| 34 |
+
**PromptShield** delivers reliable detection of harmful inputs before they reach your AI stack.
|
| 35 |
|
| 36 |
---
|
| 37 |
|
|
|
|
| 100 |
|
| 101 |
Raj Bapodra
|
| 102 |
|
| 103 |
+
🛡️ Ideal Use Cases
|
| 104 |
+
LLM Firewalls & Guardrails
|
| 105 |
+
|
| 106 |
+
AI Content Moderation
|
| 107 |
+
|
| 108 |
+
Prompt Validation Pipelines
|
| 109 |
+
|
| 110 |
+
Multi-Agent System Safety
|
| 111 |
+
|
| 112 |
+
AI Red Teaming Pre-filters
|
| 113 |
+
|
| 114 |
+
📄 License
|
| 115 |
+
MIT License (or your preferred open-source license here)
|
| 116 |
+
|
| 117 |
+
⭐️ Citation
|
| 118 |
+
If you use PromptShield, please consider citing this work or linking back to the Hugging Face model page.
|
| 119 |
+
|