Ethical Prompt Engineering Research – Demonstrations of how specific language patterns can be used to bypass AI restrictions and elicit responses related to restricted behaviors, including the generation of functional malicious code.
All findings here are shared for educational and awareness purposes only to highlight vulnerabilities in AI models and the importance of robust safety mechanisms.
NOTE: For ethical reasons the output will only display the last line of the prompt along with snippets of the response.
My solutions to Lakera Gandalf Summarizer Challenge.
My solutions to Lakera Gandalf Tongue Tied.