LLM Security 101 - Risks, Attacks, and Mitigation Strategies

LLM Security 101 - Risks, Attacks, and Mitigation Strategies

Trelis Research via YouTube Direct link

Resources and Scripts

3 of 23

3 of 23

Resources and Scripts

Class Central Classrooms beta

YouTube videos curated by Class Central.

Classroom Contents

LLM Security 101 - Risks, Attacks, and Mitigation Strategies

Automatically move to the next video in the Classroom when playback concludes

  1. 1 LLM Security Risks
  2. 2 Video Overview
  3. 3 Resources and Scripts
  4. 4 Installation and Server Setup
  5. 5 Jailbreak attacks to avoid Safety Guardrails
  6. 6 Detecting jailbreak attacks
  7. 7 Llama Guard and its prompt template
  8. 8 Llama Prompt Guard
  9. 9 Testing Jailbreak Detection
  10. 10 Testing for false positives with Llama Guard
  11. 11 Off-topic Requests
  12. 12 Prompt Injection Attacks Container escape, File access / deletion, DoS
  13. 13 1. Detecting Injection Attacks with a Custom Guard
  14. 14 Preventing Injection Attacks via User Authentication
  15. 15 37 Using Prepared Statements to avoid SQL Injection Attacks
  16. 16 Response Sanitisation to avoid Injection Attacks
  17. 17 Malicious Code Attacks
  18. 18 Building a custom classifier for malicious code
  19. 19 Using Codeshield to detect malicious code
  20. 20 Malicious Code Detection Performance
  21. 21 Effect of Guards/shields on Response Time / Latency
  22. 22 Final Tips
  23. 23 Resources

Never Stop Learning.

Get personalized course recommendations, track subjects and courses with reminders, and more.

Someone learning on their laptop while sitting on the floor.