I am recruiting PhD students & Postdocs on AI Security, LLM Agents, Code Generation research at UMD Computer Science UMD Department of Computer Science & Maryland Cybersecurity Center Maryland Cybersecurity Center (MC2) For PhD program pls mention me in your application cs.umd.edu/grad/apply. For Postdocs please email me.

LLMs may be copying training data in everyday conversations with users! In our latest work, we study how often this happens compared to humans. 👇🧵

SynthID-Text by Google DeepMind is the first large-scale LLM watermark deployment, but its behavior in adversarial scenarios is largely unexplored. In our new blogpost, we apply the recent works from SRI Lab and find that... 👇🧵

Using GCG to jailbreak Llama 3 yields only a 14% attack success rate. Is GCG hitting a wall, or is Llama 3 just safer? We found that simply replacing the generic "Sure, here is***" target prefix with our tailored prefix boosts success rates to 80%. (1/8)

🧵 Announcing Open Philanthropy's Technical AI Safety RFP! We're seeking proposals across 21 research areas to help make AI systems more trustworthy, rule-following, and aligned, even as they become more capable.

1/🔒Worried about giving your agent advanced capabilities due to prompt injection risks and rogue actions? Worry no more! Here's CaMeL: a robust defense against prompt injection attacks in LLM agents that provides formal security guarantees without modifying the underlying model!

🛠️ Still doing prompt engineering for R1 reasoning models? 🧩 Why not do some "engineering" in reasoning as well? Introducing our new paper, Effectively Controlling Reasoning Models through Thinking Intervention. 🧵[1/n]

We are starting our journey on making Gemini robust to prompt injections and in this paper we present the steps we have taken so far. A collective effort by the GDM Security & Privacy Research team spanning over > 1 year.

new paper from our work at Meta! **GPT-style language models memorize 3.6 bits per param** we compute capacity by measuring total bits memorized, using some theory from Shannon (1953) shockingly, the memorization-datasize curves look like this: ___________ / / (🧵)

I will be at ICML this year after a full long year of not attending any conference :) Happy to chat, and please don’t hesitate to reach out here, email, on Whova, or in person 🥳

Some problems can’t be rushed—they can only be done step by step, no matter how many people or processors you throw at them. We’ve scaled AI by making everything bigger and more parallel: Our models are parallel. Our scaling is parallel. Our GPUs are parallel. But what if the

Very cool thought-provoking piece! In practice, computation units are much more nuanced than what theories capture. But just trying to identify classes of problems that benefit from sequential computation (or is unsolvable without it) seems very useful!

🚨 Got a great idea for an AI + Security competition? SaTML Conference is now accepting proposals for its Competition Track! Showcase your challenge and engage the community. 👉 satml.org/call-for-compe… 🗓️ Deadline: Aug 6