Do not think about pink elephant!
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\%.
Code (0)
등록된 구현이 없습니다.
Methods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
Suppressing Pink Elephants with Direct Principle Feedback
Existing methods for controlling language models, such as RLHF and Constitutional AI, involve determining which LLM behaviors are desirable and training them into a language model. However, in many cases, it is desirable…
Language ModelingLanguage ModellingNegation: A Pink Elephant in the Large Language Models' Room?
Negations are key to determining sentence meaning, making them essential for logical reasoning. Despite their importance, negations pose a substantial challenge for large language models (LLMs) and remain underexplored. …
Language ModelingLanguage ModellingLogical ReasoningNatural Language Inference+2A simple model for pink noise from amplitude modulations
We propose a simple model for the origin of pink noise (or 1/f fluctuation) based on the beat of cooperative waves. These cooperative waves arise spontaneously in a system with synchronization, resonance, and infrared di…
The impact of physiological stress conditions on protein structure and trypsin inhibition of serine protease inhibitor Kazal type 1 (SPINK1) and its N34S variant
One of the most common mutations in the serine protease inhibitor Kazal type 1 (SPINK1) gene is the N34S variant which is strongly associated with chronic pancreatitis. Although it is assumed that N34S mutation constitut…
ElephantBook: A Semi-Automated Human-in-the-Loop System for Elephant Re-Identification
African elephants are vital to their ecosystems, but their populations are threatened by a rise in human-elephant conflict and poaching. Monitoring population dynamics is essential in conservation efforts; however, track…
Attribute