相关论文: ZombieLoad: Cross-Privilege-Boundary Data Sampling

Dropout is a common operator in deep learning, aiming to prevent overfitting by randomly dropping neurons during training. This paper introduces a new family of poisoning attacks against neural networks named DROPOUTATTACK. DROPOUTATTACK…

The Spectre and Meltdown flaws in modern microprocessors represent a new class of attacks that have been difficult to mitigate. The mitigations that have been proposed have known performance impacts. The reported magnitude of these impacts…

The rapid evolution of Large Language Models (LLMs) towards long-context reasoning and sparse architectures has pushed memory requirements far beyond the capacity of individual device HBM. While emerging supernode architectures offer…

Large language models (LLMs) have made remarkable strides in complex reasoning tasks, but their safety and robustness in reasoning processes remain underexplored. Existing attacks on LLM reasoning are constrained by specific settings or…

Guardrail models (a.k.a. safety checkers) are widely deployed to screen user inputs before they reach large language models (LLMs), serving as a primary defense against prompt injection attacks. Due to strict context constraints, these…

Modern microprocessors depend on speculative execution, creating vulnerabilities that enable transient execution attacks. Prior defenses target speculative data leakage but overlook false dependencies from partial address aliasing, where…

Prompt injection attacks aim to contaminate the input data of an LLM to mislead it into completing an attacker-chosen task instead of the intended task. In many applications and agents, the input data originates from multiple sources, with…

Memory corruption is a serious class of software vulnerabilities, which requires careful attention to be detected and removed from applications before getting exploited and harming the system users. Symbolic execution is a well-known method…

Memory allocation, though constituting only a small portion of the executed code, can have a "butterfly effect" on overall program performance, leading to significant and far-reaching impacts. Despite accounting for just approximately 5% of…

Control-flow leakage (CFL) attacks enable an attacker to expose control-flow decisions of a victim program via side-channel observations. Linearization (i.e., elimination) of secret-dependent control flow is the main countermeasure against…

Large language model (LLM) agents increasingly leverage long term memory to support persistent and autonomous task execution. However, this capability also introduces a new attack surface: memory poisoning, where adversaries can inject…

While network abrupt breakdowns due to overloads and cascading failures have been studied extensively, the critical exponents and the universality class of such phase transitions have not been discussed. Here, we study breakdowns triggered…

Register Files (RFs) are the most frequently accessed memories in a microprocessor for fast and efficient computation and control logic. Segment registers and control registers are especially critical for maintaining the CPU mode of…

Cross-workload design space exploration (DSE) is crucial in CPU architecture design. Existing DSE methods typically employ the transfer learning technique to leverage knowledge from source workloads, aiming to minimize the requirement of…

A fundamental assumption in software security is that a memory location can only be modified by processes that may write to this memory location. However, a recent study has shown that parasitic effects in DRAM can change the content of a…

LLM training at the scale of tens of thousands of GPUs now spans multiple datacenters (DC), making cross-DC collectives over long-haul links unavoidable. A critical and overlooked bottleneck arises when these collectives collide with…

A software vulnerability could be exploited without any visible symptoms. When no source code is available, although such silent program executions could cause very serious damage, the general problem of analyzing silent yet harmful…

Neural networks are seeing increased use in diverse Internet of Things (IoT) applications such as healthcare, smart homes and industrial monitoring. Their widespread use makes neural networks a lucrative target for theft. An attacker can…

Despite extensive safety measures, LLMs are vulnerable to adversarial inputs, or jailbreaks, which can elicit unsafe behaviors. In this work, we introduce bijection learning, a powerful attack algorithm which automatically fuzzes LLMs for…

Many machine learning algorithms are vulnerable to almost imperceptible perturbations of their inputs. So far it was unclear how much risk adversarial perturbations carry for the safety of real-world machine learning applications because…