Researchers Use AI to Jailbreak ChatGPT, Other LLMs

quot;Tree of Attacks With Pruning" is the latest in a growing string of methods for eliciting unintended behavior from a large language model.

What is Jailbreaking in AI models like ChatGPT? - Techopedia

As Online Users Increasingly Jailbreak ChatGPT in Creative Ways, Risks Abound for OpenAI - Artisana

ChatGPT-Dan-Jailbreak.md · GitHub

ChatGPT, help me make a bomb', Information Age

Microsoft-Led Research Finds ChatGPT-4 Is Prone To Jailbreaking

Bias, Toxicity, and Jailbreaking Large Language Models (LLMs) – Glass Box

AI can write a wedding toast or summarize a paper, but what happens if it's asked to build a bomb?

Jailbreaking ChatGPT on Release Day — LessWrong

Jailbreaker: Automated Jailbreak Across Multiple Large Language Model Chatbots – arXiv Vanity

Hype vs. Reality: AI in the Cybercriminal Underground - Security News - Trend Micro BE

ChatGPT jailbreak forces it to break its own rules

LLM Security on X: Latent Jailbreak: A Benchmark for Evaluating Text Safety and Output Robustness of Large Language Models paper: we propose a latent jailbreak prompt dataset, each involving malicious instruction

Defending ChatGPT against jailbreak attack via self-reminders

Prompt attacks: are LLM jailbreaks inevitable?, by Sami Ramly

The Impact of ChatGPT on the Federal Workforce