相关论文: Async Control: Stress-testing Asynchronous Control…

Real-world agentic tasks, unlike synchronous Markov Decision Processes (MDPs), often involve non-blocking actions with variable latencies, creating a fundamental \textit{Temporal Gap} between action initiation and completion. Existing…

Distributed systems are notoriously difficult to understand and analyze in order to assert their correction w.r.t. given properties. They often exhibit a huge number of different behaviors, as soon as the active entities (peers, agents,…

LLM agents process trusted instructions, retrieved records, and tool observations through a common generative channel. This conflates data flow with authority: an untrusted string can affect a secret-bearing response or an action proposal…

A prerequisite for safe autonomy-in-the-wild is safe testing-in-the-wild. Yet real-world autonomous tests face several unique safety challenges, both due to the possibility of causing harm during a test, as well as the risk of encountering…

In this paper, we discuss the feasibility of monitoring partially synchronous distributed systems to detect latent bugs, i.e., errors caused by concurrency and race conditions among concurrent processes. We present a monitoring framework…

Frontier language model agents can exhibit misaligned behaviors, including deception, exploiting reward hacks, and pursuing hidden objectives. To control potentially misaligned agents, we can use LLMs themselves to monitor for misbehavior.…

Large Language Models (LLMs) have been increasingly integrated into computer-use agents, which can autonomously operate tools on a user's computer to accomplish complex tasks. However, due to the inherently unstable and unpredictable nature…

For machine learning components used as part of autonomous systems (AS) in carrying out critical tasks it is crucial that assurance of the models can be maintained in the face of post-deployment changes (such as changes in the operating…

Malicious softwares or malwares for short have become a major security threat. While originating in criminal behavior, their impact are also influenced by the decisions of legitimate end users. Getting agents in the Internet, and in…

We study the problem of monitoring distributed systems where computers communicate using message passing and share an almost synchronized clock. This is a realistic scenario for networks where the speed of the monitoring is sufficiently…

The prevalent deployment of Large Language Model agents such as OpenClaw unlocks potential in real-world applications, while amplifying safety concerns. Among these concerns, the self-replication risk of LLM agents driven by objective…

Multi-agent systems, where specialized agents collaborate to solve a shared task hold great potential, from increased modularity to simulating complex environments. However, they also have a major caveat -- a single agent can cause the…

Large Language Model (LLM) agents are increasingly being deployed as conversational assistants capable of performing complex real-world tasks through tool integration. This enhanced ability to interact with external systems and process…

Large language model (LLM) agents on multi-step tasks suffer reasoning degradation, looping, drift, stuck states, at rates up to 30% on hard tasks. Current solutions include hard step limits (abrupt) or LLM-as-judge monitoring (10-15%…

Driven by the rapid development of Large Language Models (LLMs), LLM-based agents have been developed to handle various real-world applications, including finance, healthcare, and shopping, etc. It is crucial to ensure the reliability and…

Maintaining the security of control systems in the presence of integrity attacks is a significant challenge. In literature, several possible attacks against control systems have been formulated including replay, false data injection, and…

As AI systems are increasingly used to conduct research autonomously, misaligned systems could introduce subtle flaws that produce misleading results while evading detection. We introduce Auditing Sabotage Bench, a benchmark for evaluating…

Current LLM safety defenses fail under decomposition attacks, where a malicious goal is decomposed into benign subtasks that circumvent refusals. The challenge lies in the existing shallow safety alignment techniques: they only detect harm…

This paper studies the synchronization of a multi-agent system where the agents are coupled through heterogeneous controller gains. Synchronization refers to the situation where all the agents in a group have a common velocity direction. We…

Monitoring software systems at runtime is key for understanding workloads, debugging, and self-adaptation. It typically involves collecting and storing observable software data, which can be analyzed online or offline. Despite the…