Alerts
Events
DCR
Explore Cyware Products
Alerts
Events
DCR
Go to listing page
ChatGPT can be Manipulated Using Hexadecimal Encoding
Malware and Vulnerabilities
October 31, 2024
Dark Reading
This technique leverages the model's ability to process multiple input formats and its focus on step-by-step instructions. By encoding malicious instructions in hexadecimal format, the model is tricked into decoding and executing them.
Read More
OpenAI
OpenAI GPT 4o
GPT-4o
Large Language Model (LLM)
Compartmentalized Task Execution
Publisher
Previous
Chenlun’s Evolving Phishing Tactics Target Trusted Bran ...
Identity Theft, Fraud, Scams
Next
Massive PSAUX Ransomware Attack Targets 22,000 CyberPan ...
Malware and Vulnerabilities