Beyond Guardrails: Defending LLMs Against Sophisticated Attacks 1rj3s

The Data Exchange with Ben Lorica

22/05/2025

Jason Martin is an AI Security Researcher at HiddenLayer. This episode explores “policy...

Jason Martin is an AI Security Researcher at HiddenLayer. This episode explores “policy puppetry,” a universal attack technique bying safety features in all major language models using structured formats like XML or JSON.
Subscribe to the Gradient Flow Newsletter 📩  https://gradientflow.substack.com/
Subscribe: Apple · Spotify · Overcast · Pocket Casts · AntennaPod · Podcast Addict · Amazon ·  RSS.

Detailed show notes - with links to many references - can be found on The Data Exchange web site.

The Highly Uncertain Future of OpenAI’s Dominance 15 días 54:07 How a Public-Benefit Startup Plans to Make Open Source the Default for Serious AI 8 días 48:45 From Vibe Coding to Autonomous Agents 1 día 51:16 Navigating the Generative AI Maze in Business 29 días 49:35 The Practical Realities of AI Development 1 mes 37:31 Ver más en APP Comentarios del episodio 5zu1y