Do not think about pink elephant!
This addresses security and ethical issues for users and providers of large models by exposing and countering a specific vulnerability, though it is incremental as it builds on known cognitive biases.
The paper tackles the vulnerability of large models like Stable Diffusion and DALL-E3 to the 'white bear phenomenon', similar to human cognitive biases, by analyzing their representation space and proposing a prompt-based attack method that generates prohibited content, with defense strategies mitigating attacks by up to 48.22%.
Large Models (LMs) have heightened expectations for the potential of general AI as they are akin to human intelligence. This paper shows that recent large models such as Stable Diffusion and DALL-E3 also share the vulnerability of human intelligence, namely the "white bear phenomenon". We investigate the causes of the white bear phenomenon by analyzing their representation space. Based on this analysis, we propose a simple prompt-based attack method, which generates figures prohibited by the LM provider's policy. To counter these attacks, we introduce prompt-based defense strategies inspired by cognitive therapy techniques, successfully mitigating attacks by up to 48.22\%.