Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2411.14738 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915030992158720 |
|---|---|
| author | Liang, Jiashuo Li, Guancheng Yu, Yang |
| author_facet | Liang, Jiashuo Li, Guancheng Yu, Yang |
| contents | Large language models (LLMs) have been widely adopted in applications such as automated content generation and even critical decision-making systems. However, the risk of prompt injection allows for potential manipulation of LLM outputs. While numerous attack methods have been documented, achieving full control over these outputs remains challenging, often requiring experienced attackers to make multiple attempts and depending heavily on the prompt context. Recent advancements in gradient-based white-box attack techniques have shown promise in tasks like jailbreaks and system prompt leaks. Our research generalizes gradient-based attacks to find a trigger that is (1) Universal: effective irrespective of the target output; (2) Context-Independent: robust across diverse prompt contexts; and (3) Precise Output: capable of manipulating LLM inputs to yield any specified output with high accuracy. We propose a novel method to efficiently discover such triggers and assess the effectiveness of the proposed attack. Furthermore, we discuss the substantial threats posed by such attacks to LLM-based applications, highlighting the potential for adversaries to taking over the decisions and actions made by AI agents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_14738 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Universal and Context-Independent Triggers for Precise Control of LLM Outputs Liang, Jiashuo Li, Guancheng Yu, Yang Computation and Language Artificial Intelligence Cryptography and Security Large language models (LLMs) have been widely adopted in applications such as automated content generation and even critical decision-making systems. However, the risk of prompt injection allows for potential manipulation of LLM outputs. While numerous attack methods have been documented, achieving full control over these outputs remains challenging, often requiring experienced attackers to make multiple attempts and depending heavily on the prompt context. Recent advancements in gradient-based white-box attack techniques have shown promise in tasks like jailbreaks and system prompt leaks. Our research generalizes gradient-based attacks to find a trigger that is (1) Universal: effective irrespective of the target output; (2) Context-Independent: robust across diverse prompt contexts; and (3) Precise Output: capable of manipulating LLM inputs to yield any specified output with high accuracy. We propose a novel method to efficiently discover such triggers and assess the effectiveness of the proposed attack. Furthermore, we discuss the substantial threats posed by such attacks to LLM-based applications, highlighting the potential for adversaries to taking over the decisions and actions made by AI agents. |
| title | Universal and Context-Independent Triggers for Precise Control of LLM Outputs |
| topic | Computation and Language Artificial Intelligence Cryptography and Security |
| url | https://arxiv.org/abs/2411.14738 |