Learn PromptingIntroduction
Prompt hacking is a term used to describe a type of attack that exploits the vulnerabilities of LLM , by manipulating their inputs or prompts. Unlike traditional hacking, which typically exploits software vulnerabilities, prompt hacking relies on carefully crafting prompts to deceive the LLM into performing unintended actions.
We will cover three types of prompt hacking: prompt injection, prompt leaking, and jailbreaking. Prompt injection involves adding malicious or unintended content to a prompt to hijack the language model's output. Prompt leaking and jailbreaking are effectively subsets of this: Prompt leaking involves extracting sensitive or confidential information from the LLM's responses, while jailbreaking involves bypassing safety and moderation features. We will also discuss specific offensive techniques as well as defensive techniques.
To protect against prompt hacking, defensive measures must be taken. These include implementing prompt based defenses, regularly monitoring the LLM's behavior and outputs for unusual activity, and using fine tuning or other techniques. Overall, prompt hacking is a growing concern for the security of LLMs, and it is essential to remain vigilant and take proactive steps to protect against these types of attacks.
Sander Schulhoff
Sander Schulhoff is the CEO of HackAPrompt and Learn Prompting. He created the first Prompt Engineering guide on the internet, two months before ChatGPT was released, which has taught 3 million people how to prompt ChatGPT. He also partnered with OpenAI to run the first AI Red Teaming competition, HackAPrompt, which was 2x larger than the White House's subsequent AI Red Teaming competition. Today, HackAPrompt partners with the Frontier AI labs to produce research that makes their models more secure. Sander's background is in Natural Language Processing and deep reinforcement learning. He recently led the team behind The Prompt Report, the most comprehensive study of prompt engineering ever done. This 76-page survey, co-authored with OpenAI, Microsoft, Google, Princeton, Stanford, and other leading institutions, analyzed 1,500+ academic papers and covered 200+ prompting techniques.