keeganmccallum / llm-guard

The Security Toolkit for LLM Interactions

LLM Guard - The Security Toolkit for LLM Interactions

LLM-Guard is a comprehensive tool designed to fortify the security of Large Language Models (LLMs).

Documentation | Demo

❤️ Proudly developed by the Laiyer.ai team.

What is LLM Guard?

By offering sanitization, detection of harmful language, prevention of data leakage, and resistance against prompt injection attacks, LLM-Guard ensures that your interactions with LLMs remain safe and secure.

Installation

Begin your journey with LLM Guard by downloading the package and acquiring the en_core_web_trf spaCy model (essential for the Anonymize scanner):

pip install llm-guard
python -m spacy download en_core_web_trf

Getting Started

Important Notes:

• LLM Guard is designed for easy integration and deployment in production environments. While it's ready to use out-of-the-box, please be informed that we're constantly improving and updating the repository.
• Base functionality requires a limited number of libraries. As you explore more advanced features, necessary libraries will be automatically installed.
• Ensure you're using Python version 3.8.1 or higher. Confirm with: python --version.
• Library installation issues? Consider upgrading pip: python -m pip install --upgrade pip.

Examples:

• Get started with ChatGPT and LLM Guard.

Supported scanners

Prompt scanners

• Anonymize
• BanSubstrings
• BanTopics
• Code
• PromptInjection
• Secrets
• Sentiment
• TokenLimit
• Toxicity

Output scanners

• BanSubstrings
• BanTopics
• Bias
• Code
• Deanonymize
• MaliciousURLs
• NoRefusal
• Refutation
• Regex
• Relevance
• Sensitive
• Sentiment
• Toxicity

Roadmap

General:

• Introduce support of GPU
• Improve documentation by showing use-cases, benchmarks, etc
• Hosted version of LLM Guard
• Text statistics to provide on prompt and output
• Support more languages
• Accept multiple outputs instead of one to compare
• Support streaming mode

Prompt Scanner:

• Integrate with Perspective API for Toxicity scanner
• Develop language restricting scanner

Output Scanner:

• Develop output scanners for the format (e.g. max length, correct JSON, XML, etc)
• Develop factual consistency scanner
• Develop libraries hallucination scanner
• Develop libraries licenses scanner

Contributing

Got ideas, feedback, or wish to contribute? We'd love to hear from you! Email us.

For detailed guidelines on contributions, kindly refer to our contribution guide.