Rebuff | Open Source Framework To Detect and Protect Prompt Injection | LLM
Rebuff is an open-source framework designed to detect and protect against prompt injection attacks in Language Learning Model (LLM) applications.
In this video, I will show you how you can utilize Rebuff in your AI applications to protect from malicious attacks / inputs.
Rebuff uses multiple layers of defense to protect LLM applications:
Heuristics: Rebuff incorporates heuristics to filter out potentially malicious input before it reaches the LLM.
LLM-based detection: Rebuff uses a dedicated LLM to analyse incoming prompts and identify potential attacks.
VectorDB: Rebuff stores embeddings of previous attacks in a vector database, enabling it to recognize and prevent similar attacks in the future.
Canary tokens: Rebuff adds canary tokens to prompts to detect leakages, which then allows the framework to store embeddings about the incoming prompt in the vector database and prevent future attacks.
👉🏼 Links:
langchain rebuff blog: https://blog.langchain.dev/rebuff/
rebuff website: https://rebuff.ai/
langchain rebuff documentation: https://python.langchain.com/en/lates...
☕ Buy me a Coffee: https://ko-fi.com/datasciencebasics
🔗 Other videos you might find helpful:
⛓️ Langflow: • ⛓️ langflow | UI For 🦜️🔗 LangChain
🔥 PrivateGPT: • PrivateGPT 🔥 | No Data Leaks | 100% F...
🦜️🔗 LangChain playlist: • LangChain
🐍🌐💻 Python Virtual Env: • Python Virtual Environment
🤝 Connect with me:
🔴 Youtube: https://www.youtube.com/@datascienceb...
👔 LinkedIn: / sudarshan-koirala
🐦 Twitter: / mesudarshan
#langchain #llm #gpt #promptinjection #openai