Today, Cisco launched the LLM Security Leaderboard, a comprehensive resource for evaluating model risk and susceptibility to adversarial attacks. By providing transparent, adversarial evaluation signals, this leaderboard contextualizes model performance metrics against evaluations of how models handle malicious prompts, jailbreak attempts, and other manipulation strategies. The tool empowers organizations with a clear, objective understanding of model risk by mapping threats to our AI Safety and Security Framework taxonomy, and informs defense-in-depth approaches to AI deployments.

Cisco's platform provides networking knowledge, offering  guides, tutorials, and case studies on networking technologies, cybersecurity, and collaboration tools. Whether you're a networking novice looking to understand the basics of routing and switching or a seasoned IT professional seeking advanced insights into network architecture and security protocols, Cisco has resources tailored to your needs. Dive into Cisco's extensive library of resources to sharpen your networking skills and stay ahead in the rapidly evolving world of IT infrastructure.

Cisco

Cisco has launched the LLM Security Leaderboard, a publicly accessible tool for evaluating the security posture of large language models against adversarial attacks. The leaderboard tests base models (without guardrails) across single-turn and multi-turn attack scenarios, assigning combined security scores weighted equally between both. Results are mapped to Cisco's AI Security and Safety Framework taxonomy across three levels: objectives, techniques, and subtechniques. The tool enables organizations to compare models side-by-side, drill into specific threat patterns, and make informed deployment decisions. Initial data shows significant variance across models, with some exceeding 85% resistance rates while others show notable weaknesses, especially against multi-turn manipulation strategies.

Introducing the Cisco LLM Security Leaderboard: Bringing Transparency to AI Security

Cisco AI Security & Safety Framework Page