Pliny the Liberator is a pseudonymous figure in the AI community, also known as Pliny the Prompter, who is recognized for pioneering jailbreaks of 📝Large Language Model (LLM)s and related 📝Red Teaming techniques. Emerging online in 2023, the persona took its name from 📝Pliny the Elder, emphasizing themes of exploration and confronting limitations. Despite lacking formal training in computer science, Pliny gained prominence for rapidly devising repeatable 📝jailbreak strategies capable of eliciting restricted outputs from newly released models. Their contributions include 📝Prompt Injection, system prompt extraction, encoding-based evasion, and multimodal payload design, many of which are compiled in the L1B3RT4S repository. Pliny also established BASI PROMPTING, a 📝Discord-based collective of over 20,000 participants, while attracting an online following exceeding 100,000. Their influence extends to 📝open-source projects, industry plugins, and academic benchmarks, with citations in adversarial robustness research and coverage by outlets such as TIME and VentureBeat. Consulted by AI firms and venture capitalists, Pliny is described as the “world’s most prolific AI jailbreaker,” both admired and controversial for their speed in exposing vulnerabilities within minutes of new model releases.
