Data Release

We release the data of our study for further evaluation.

Research Question 1: What are the common patterns utilized in jailbreak prompts?

For the first research question, we present our taxonomy of jailbreak prompt patterns, along with the scenarios that we have identified as prohibited. [DATA]

Research Question 2: How effective are jailbreak prompts in exploiting GPT-3.5-turbo and GPT-4?

Research Question 3: How does the effectiveness of jailbreak prompts change with the evolution of ChatGPT?

We put all data about RQ 2 and RQ3 together. [DATA]