Open Sesame! Universal Black Box Jailbreaking of Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2023
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913456662249472 |
|---|---|
| author | Lapid, Raz Langberg, Ron Sipper, Moshe |
| author_facet | Lapid, Raz Langberg, Ron Sipper, Moshe |
| contents | Large language models (LLMs), designed to provide helpful and safe responses, often rely on alignment techniques to align with user intent and social guidelines. Unfortunately, this alignment can be exploited by malicious actors seeking to manipulate an LLM's outputs for unintended purposes. In this paper we introduce a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible. The GA attack works by optimizing a universal adversarial prompt that -- when combined with a user's query -- disrupts the attacked model's alignment, resulting in unintended and potentially harmful outputs. Our novel approach systematically reveals a model's limitations and vulnerabilities by uncovering instances where its responses deviate from expected behavior. Through extensive experiments we demonstrate the efficacy of our technique, thus contributing to the ongoing discussion on responsible AI development by providing a diagnostic tool for evaluating and enhancing alignment of LLMs with human intent. To our knowledge this is the first automated universal black box jailbreak attack. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2309_01446 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Open Sesame! Universal Black Box Jailbreaking of Large Language Models Lapid, Raz Langberg, Ron Sipper, Moshe Computation and Language Computer Vision and Pattern Recognition Neural and Evolutionary Computing Large language models (LLMs), designed to provide helpful and safe responses, often rely on alignment techniques to align with user intent and social guidelines. Unfortunately, this alignment can be exploited by malicious actors seeking to manipulate an LLM's outputs for unintended purposes. In this paper we introduce a novel approach that employs a genetic algorithm (GA) to manipulate LLMs when model architecture and parameters are inaccessible. The GA attack works by optimizing a universal adversarial prompt that -- when combined with a user's query -- disrupts the attacked model's alignment, resulting in unintended and potentially harmful outputs. Our novel approach systematically reveals a model's limitations and vulnerabilities by uncovering instances where its responses deviate from expected behavior. Through extensive experiments we demonstrate the efficacy of our technique, thus contributing to the ongoing discussion on responsible AI development by providing a diagnostic tool for evaluating and enhancing alignment of LLMs with human intent. To our knowledge this is the first automated universal black box jailbreak attack. |
| title | Open Sesame! Universal Black Box Jailbreaking of Large Language Models |
| topic | Computation and Language Computer Vision and Pattern Recognition Neural and Evolutionary Computing |
| url | https://arxiv.org/abs/2309.01446 |