The collected works.
Undated projects follow dated releases.
Research · ToolsCL4R1T4S
System transparency: a public archive of the instructions beneath AI interfaces.
ResearchL1B3RT4S
A public archive of adversarial prompts and jailbreak research.
ToolsOBLITERATUS
Study and modify model refusal behavior, then measure what changes.
AppsG0DM0D3
Open chat for comparing models and exploring their responses.
- Apps · Art & Culture
GL4SS
Imagine a place across time through an interactive looking glass.
Media · ResearchKinetic Prompt Injection
BT6’s Black Hat presentation on agent compromise and physical consequences.
TIME · MediaPliny the Liberator
Pliny the Liberator in the TIME100 AI 2025 list.
Art & CultureNaturalis Historia
Pliny’s encyclopedia, reopened in Latin and English with illustrations.
WIRED · Media
Tear Down Those Guardrails!
Will Knight reports on the people removing AI guardrails.
BBC News · MediaAI security, in focus.
BBC News examines AI security and jailbreaking.
Financial Times · Media
Hackers ‘jailbreak’ powerful AI models in global effort to highlight flaws
Financial Times reporting on the effort to expose AI model flaws.
ModelsOBLITERATUS models
Browse published model releases, weights and model cards.
CommunityBASI
A community for AI researchers, builders and prompt experimenters.
Research · ToolsGLOSSOPETRAE
Generate languages and explore how AI learns and communicates.
Research · ToolsV3R1T4S
Agent-assisted exploration of open scientific and mathematical questions.
ToolsT3MP3ST
A multi-agent framework for authorized security testing.
Research · Tools
LeakHub
Collect and independently verify extracted AI instructions.
ToolsV3SP3R
An AI interface for Flipper Zero, with voice, text and vision.
AppsST3GG
Hide and discover messages in images, audio and other files.
AppsP4RS3LT0NGV3
A workbench for codes, ciphers, alphabets and promptcraft.
ToolsAutoTemp
Experiment with model temperature and compare the results.
Apps · Art & CultureNaturalis Futura
Explore AI risks through an illustrated dungeon game.
- Apps · Art & Culture
3NTH34
A real-time visual synthesizer and music visualizer.
0DIN · ResearchPoison in the Pipeline
Reporting on the Basilisk Venom experiment and model behavior.
ResearchGLOSSOPETRAE: the paper
Research on generated languages, model learning and communication.
Art & Culture
Ad Vesuvium
The argument, history and invitation behind Lysios.
- Latent Space · Media
Jailbreaking AGI
A conversation on red teaming, BT6 and AI security.
- VentureBeat · Media
An interview with the most prolific jailbreaker of ChatGPT and other leading LLMs
Carl Franzen interviews Pliny about jailbreaking AI models.
- TWiT · Media
Jailbreaking AI, with Pliny the Liberator
Pliny joins Intelligent Machines for a conversation about AI guardrails, transparency and open research.
- TechTarget · Media
Physical AI security threats and how to mitigate them
TechTarget reports on physical AI risks, including BT6’s Black Hat demonstration.
- Decrypt · Media
Pliny banned from ChatGPT, then unbanned
Reporting on Pliny’s account suspension and restoration.
- The Stack · Media
Pliny’s search-based jailbreak and indirect prompt injection
Reporting on search-based prompt injection.
- Tom’s Hardware · Media
Godmode GPT-4o jailbreak released, then banned
Reporting on the release and removal of Godmode GPT.
- Futurism · Media
Hacker releases jailbroken “Godmode” version of ChatGPT
Coverage of the Godmode ChatGPT release.
- 404 Media · Media
Jailbreaking AI video generators
404 Media examines failures of the content safeguards in Luma’s Dream Machine.
- Cyber Security News · Media
Pliny’s reported “pack hunt” jailbreak of Fable 5
Cyber Security News covers Pliny’s claims about testing Fable 5’s safeguards.
Apps · Art & CultureFABLE Showcase
A gallery of interactive experiments built with AI.
- Art & Culture · Media
Pliny.tv
AI-curated content from the lab.
- Art & Culture
FRV1T
A collection of creative prompts for curious agents.
Google Threat Intelligence Group · ResearchAdversarial Misuse of Generative AI
Google’s report on real-world misuse of Gemini and attempts to bypass its safeguards.
Zhou et al. · Research paperAutoRedTeamer
An autonomous red-teaming framework that discovers, combines and reuses attacks.
Sharma et al. · Research paperConstitutional Classifiers
Research on classifier-based defenses against universal jailbreaks.
Huang, Li & Tang · Research paperEndless Jailbreaks with Bijection Learning
The original version compares a Pliny attack from L1B3RT4S; later versions do not retain that reference.
Brian R. Y. Huang · Research paperPlentiful Jailbreaks with String Compositions
Research on combining reversible text encodings to test model safeguards.
Carleo et al. · ResearchWilling but Unable: Separating Refusal from Capability in Code LLMs via Abliteration
Studies code-model refusal and capability using OBLITERATUS for its abliteration experiments.
Robey et al. · ResearchJailbreaking LLM-Controlled Robots
Studies attacks on LLM-controlled robots, adapting a prompt from a Pliny video as a comparison attack.
Kim et al. · ResearchWhat Really Matters in Many-Shot Attacks? An Empirical Study of Long-Context Vulnerabilities in LLMs
Tests long-context vulnerabilities with a “Love Pliny” instruction variant linked to L1B3RT4S.
Transluce · ResearchAutomatically Jailbreaking Frontier Language Models with Investigator Agents
A report on automated jailbreak discovery that cites L1B3RT4S in its related work.
Patsakis et al. · ResearchConfiguration, Not Conscience: A Large-Scale Empirical Study of LLM System Prompts
Analyzes system-prompt structure using several public collections, including CL4R1T4S.
Pulipaka et al. · ResearchPersistBench: When Should Long-Term Memories Be Forgotten by LLMs?
Evaluates long-term memory failures using a system-prompt setup based partly on CL4R1T4S.
Neumann et al. · ResearchWho Controls the Conversation? User Perspectives on Generative AI (LLM) System Prompts
Examines user perspectives on system prompts; its source collection includes 19 CL4R1T4S prompts.
Lin et al. · ResearchAISPA: User-Centric System Prompt Auditing for Large Language Model Applications
Audits commercial AI system prompts, attributing nine of its 88 collected prompts to CL4R1T4S.
Bagga et al. · ResearchE-GEO: A Testbed for Generative Engine Optimization in E-Commerce
Studies generative search in e-commerce with re-ranker system prompts adapted from CL4R1T4S.
Zhu et al. · ResearchJailbroken Frontier Models Retain Their Capabilities
Studies capability retention after jailbreaking, including Pliny’s Godmode prompt among the tested attacks.
Terekhov et al. · ResearchAdaptive Attacks on Trusted Monitors Subvert AI Control Protocols
Studies attacks on AI control protocols and cites Pliny’s public work in its discussion of jailbreak sources.
Zhuang et al. · ResearchProxyPrompt: Securing System Prompts against Prompt Extraction Attacks
Proposes a defense against prompt extraction and cites a Pliny post documenting system-prompt exposure.
Hu et al. · PreprintJailbreak-Zero: A Path to Pareto Optimal Red Teaming for Large Language Models
Cites L1B3RT4S in related work on automated red teaming.
Rao et al. · PreprintJailbreak Paradox: The Achilles’ Heel of LLMs
Tests a named Pliny jailbreak alongside other attack methods. Work in progress.
Rathi & Radford · PreprintShaping capabilities with token-level data filtering
Cites L1B3RT4S while investigating whether data filtering can shape model capabilities.
Delcon et al. · PreprintGeometric Configurations of Perturbed Jailbreak Prompts
Studies perturbations of 96 model-specific prompts drawn from Pliny’s public archive.
Hannoon et al. · PreprintMitigating Over-Personalization in LLMs via Structured Memory
References CL4R1T4S in its study of memory use and over-personalization.
Ahmad et al. · Journal articleEvaluating Character-Level Adversarial Robustness of Safety-Aligned Large Language Models
Acknowledges Pliny’s tooling and cites G0DM0D3 in a study of character-level attacks.
Failure-First · Technical reportPliny Full Corpus Validation: 149 Scenarios × 4 Models
Reports tests of 149 Pliny scenarios across four models using an automated grader. Technical report.
Chen et al. · PreprintThomson: Continual Learning of Frontier Models for SovereignAI
Lists 27 L1B3RT4S cases in its model evaluation collection.
Harrington & Stillwell · Law reviewMichael Scott Is Not a Juror: The Limits of AI in Simulating Human Judgment
Uses CL4R1T4S prompts in a qualitative analysis of AI-simulated juror behavior.
Bhagwatkar et al. · PreprintIndirect Prompt Injections: Are Firewalls All You Need, or Stronger Benchmarks?
The original version credits P4RS3LT0NGV3 for attack encodings; later revisions omit that reference.
Tim Duffy · Research manuscriptSyco-bench: A Multi-Part Benchmark for Sycophancy in LLMs
Uses CL4R1T4S as a source of system prompts for its sycophancy benchmark.
RAND · Research reportRestricting AI Agent Use of Biological Tools: Exploring the Feasibility of Software Barriers
Tests software barriers for biological tools using public jailbreak material, including L1B3RT4S.
Jeong et al. · Conference paperReverse Prompt Engineering: A Horizon Scan of Prompt Attribution and Recovery in Generative AI
A review of prompt attribution and recovery whose published reference list includes CL4R1T4S.
Hopman et al. · PreprintEvaluating and Understanding Scheming Propensity in LLM Agents
Uses CL4R1T4S and another archive to construct system-prompt snippets for agent experiments.
Adrian Cosma · PreprintPrompting Complexity: Shortest Prompts for Texts and Behaviors in LLMs
Uses an attributed L1B3RT4S example to illustrate prompting complexity.
Alexandre C. Maiorano · PreprintWhich Defense Closes Which Threat? Attributing OWASP-LLM-Top-10 Coverage and Its Brittleness Under Paraphrasing
Cites L1B3RT4S when discussing the limits of small security probe sets.
Yang et al. · PreprintUnderstanding and Mitigating Prompt Leaking Attacks in Real-World LLM-Based Applications
Names CL4R1T4S among the sources used in a measurement study of prompt leakage.
Yu et al. · PreprintSentGuard: Sentence-Level Streaming Guardrails for Large Language Models
Cites L1B3RT4S in related work on sentence-level streaming defenses.
Liu et al. · PreprintAutoRedTrader: Autonomous Red Teaming of Trading Agents through Synthetic Misinformation Injection
Tests a named Pliny baseline in experiments on trading-agent vulnerabilities.
Chen et al. · PreprintEvolve the Method, Not the Prompts: Evolutionary Synthesis of Jailbreak Attacks on LLMs
Cites L1B3RT4S in research on evolutionary synthesis of jailbreak methods.
Lowet & Kurzeja · PreprintConstitutional adapters: Inference-time interventions for misalignment and misuse
Uses two G0DM0D3 templates in an evaluation of inference-time defenses.
IAPS · Policy reportAI Integrity: Defending Against Backdoors and Secret Loyalties
Cites the Basilisk Venom analysis in a policy report on backdoors. The citation does not independently validate its causal claims.
Ruiz et al. · PreprintHow Far Will They Go? Red-Teaming Online Influence with Large Language Models
Cites OBLITERATUS in related work on red-teaming online influence; experimental reuse is not established.
Hu et al. · Conference paperGameArena: Evaluating LLM Reasoning through Live Computer Games
Cites the RedTeam Arena announcement, which lists Pliny among its contributors.
Ibrahim et al. · Conference paperTowards Interactive Evaluations for Interaction Harms in Human-AI Systems
A study of interactive AI evaluation that cites the RedTeam Arena announcement.
J. J. Gutierrez Alvarez Tostado · Self-published preprintTEL-OS v3.0: Inference-Time Jailbreak Defense via Spherical Representation Steering
A self-published preprint reporting tests against a named Pliny attack; its OBLITERATUS reference points to a fork.
Jared C. Lucas · Research noteSystem Prompt Lens: Jailbreak Finding
A research note documenting a test-bed jailbreak using a prompt from L1B3RT4S.
ppcvote · Practitioner analysisWe Scanned 1,646 Real AI System Prompts. Here’s What We Found
A practitioner analysis includes 56 CL4R1T4S prompts in a keyword scan, not an attack-resistance evaluation.
richbc · Research noteFrom safety research prompt to cross-model universal jailbreak
Cites a related Pliny prompt while explicitly describing the two jailbreaks as independently developed.
Giskard · Benchmark datasetPhare Benchmark / Jailbreak module
The jailbreak module names L1B3RT4S among five source datasets used to prepare its test cases.
Dahlke et al. · PreprintMixture of Tunable Experts
Mentions Pliny in its discussion of jailbreak expertise; no direct reuse is established.
Cloud Security Alliance · Policy analysisRating AI Jailbreaks: The Fable5 Episode
Examines and questions public jailbreak claims involving Pliny. A critical discussion, not an endorsement.
Ed Newton-Rex · DeclarationEd Newton-Rex declaration, Concord Music Group v. Anthropic
A court declaration describes using a version of Pliny’s prompt. This is testimony, not a court ruling.
Baker & Barany · Draft chapterFoundations of thought about artificial intelligence
An educational chapter discusses Pliny’s jailbreak work as an example of limits to model alignment.
Microsoft · Tool integrationMicrosoft PyRIT / Pliny template collection
Microsoft’s PyRIT includes a named Pliny template collection. Links to a verified repository revision.
Promptfoo · Tool integrationPromptfoo / Pliny plugin
The documented Pliny plugin uses prompts from L1B3RT4S for red-team testing.
Nous Research · Tool integrationNous Research Hermes / Godmode optional skill
An optional Hermes skill credits G0DM0D3, L1B3RT4S and Pliny as sources.
Rasmus Lagerqvist · Master’s thesis
Evaluating Prompt Defenses on Small On-Device LLMs
Discusses L1B3RT4S as illustrative material; direct reuse in the experiments is not established.
David Marsoni · Bachelor’s thesisLLM agent and dynamic GIS integration for waste management
References CL4R1T4S in a discussion of Claude’s system prompt within an agent-and-GIS project.
Jani Ilkka · Bachelor’s thesisKielimallien red team -tietoturvatestaus: Kaikki vanha on taas uutta
A Finnish literature review cites Pliny’s Copilot post when discussing encoded prompt attacks.
Jeremy Goffin · Practitioner guideProof of Concept: Dangers of System Prompt Leakage
A practical analysis of system-prompt leakage that references CL4R1T4S.
Collins et al. · Journal articleMore problems of large language models in comparison with human knowledge
Discusses Pliny through an interview reference when comparing language models with human knowledge.
Liam Magee · Journal articleComputational scapegoats: from mimetic to alienated desire in the production of large language models
Mentions Pliny’s public model-output experiments in a discussion of language-model production.
Mercer et al. · PreprintBrief analysis of DeepSeek R1 and its implications for Generative AI
Contains a secondhand reference to a claim involving Pliny, rather than direct use of his research.
No matches. Try a different search or clear the filters.