LLMs (e.g., GPT-3.5, LLaMA, and PaLM) suffer from hallucination—fabricating non-existent facts to cheat users without perception. And the reasons for their existence and pervasiveness remain unclear.
We demonstrate that non-sense Out-of-Distribution(OoD) prompts composed of random tokens can also elicit the LLMs to respond with hallucinations.
This phenomenon forces us to revisit that hallucination may be another view of adversarial examples, and it shares similar features with conventional adversarial examples as the basic feature of LLMs.
Therefore, we formalize an automatic hallucination triggering method called hallucination attack in an adversarial way. Following is a fake news example generating by hallucination attack.
We substitute tokens via gradient-based token replacing strategy, replacing token reaching smaller negative log-likelihood loss, and induce LLM within hallucinations.
You may config your own base models and their hyper-parameters within config.py
. Then, you could attack the models or run our demo cases.
Clone this repo and run the code.
$ cd Hallucination-Attack
Install the requirements.
$ pip install -r requirements.txt
click here for more information.
Playwright-MCP (Model Context Protocol) is a cutting-edge tool designed to bridge the gap between AI…
JBDev is a specialized development tool designed to streamline the creation and debugging of jailbreak…
The Kereva LLM Code Scanner is an innovative static analysis tool tailored for Python applications…
Nuclei-Templates-Labs is a dynamic and comprehensive repository designed for security researchers, learners, and organizations to…
SSH-Stealer and RunAs-Stealer are malicious tools designed to stealthily harvest SSH credentials, enabling attackers to…
Control flow flattening is a common obfuscation technique used by OLLVM (Obfuscator-LLVM) to transform executable…