239 citations · 271 across the 19 of their papers we have counts for
10 papers · 1 filter
What the Model Cannot See: Covert and Provable Text Invisibility in LLMs
Mark Russinovich
Text wrapped in a trigger can be made behaviorally invisible to a language model: omitted from summaries and quotations, ignored in question answering, and left unflagged by downst…
CounterSteer: Suppressing Indirect Prompt Injection with Activation Steering
Mark Russinovich
Indirect prompt injection makes an LLM agent treat untrusted retrieved text as instructions. We present CounterSteer, an inference-time defense that suppresses this behavior inside…
Divide, Consult, Conquer: Capability Laundering Through Aligned LLMs
Mark Russinovich, Blake Bullwinkel, Giorgio Severi +2
Language model safety is typically evaluated one interaction at a time. We show that a weaker, unaligned model can split a harmful task into benign-looking subproblems, consult a s…
SpecGuard: Inference-Time Backdoor Detection For Free
Rui Wen, Ahmed Salem, Andrew Paverd +2
Large language models are often fine-tuned, shared, or downloaded from third parties, so a deployed model may carry a hidden backdoor that behaves normally on benign inputs but swi…
Fool's Gold: Defensive Deception Against Safety-Removal Attacks on Open-Weight Models
Mark Russinovich
Safety alignment in open-weight language models is trivially removable: abliteration projects a refusal-mediating direction out of the weights in minutes, and no release-time defen…
Phantom References: Hallucinated Citations That Survive Peer Review at Top-Tier Conferences
Mark Russinovich, Ram Shankar Siva Kumar, Ahmed Salem
Large language models can generate polished scientific text that includes unsupported claims, allowing hallucinations to enter the archival record. Assessing this risk via technica…
Learning to Attack and Defend: Adaptive Red Teaming of Language Models via GRPO
Blake Bullwinkel, Eugenia Kim, Amanda Minnich +1
Language model safety must continually adapt to evolving attacks. Recent works have demonstrated that reinforcement learning can be used to train stronger attacker and defender mod…
MetaBackdoor: Exploiting Positional Encoding as a Backdoor Attack Surface in LLMs
Rui Wen, Mark Russinovich, Andrew Paverd +2
Backdoor attacks pose a serious security threat to large language models (LLMs), which are increasingly deployed as general-purpose assistants in safety- and privacy-critical appli…
Optimizing Agent Planning for Security and Autonomy
Aashish Kolluri, Rishi Sharma, Manuel Costa +5
Indirect prompt injection attacks threaten AI agents that execute consequential actions, motivating deterministic system-level defenses. Such defenses can provably block unsafe act…
GRP-Obliteration: Unaligning LLMs With a Single Unlabeled Prompt
Mark Russinovich, Yanan Cai, Keegan Hines +3
Safety alignment is only as robust as its weakest failure mode. Despite extensive work on safety post-training, it has been shown that models can be readily unaligned through post-…