About the poster
The poster introduces a study of offensive and defensive methods for LLM-powered bot detection. It connects adversarial content rewriting and prompt injection with an ensemble defence approach.
The source document includes both a short research paper and the poster.