Related papers: Willful Disobedience: Automatically Detecting Fail…
The safety of autonomous AI agents is increasingly recognized as a critical open problem. As agents transition from passive text generators to active actors capable of executing shell commands, modifying files, calling APIs, and browsing…
Agentic AI shifts LLM serving from isolated prompt-generation requests to stateful, multi-turn executions that repeatedly invoke the model, call tools, and grow context over time. This paper characterizes ReAct-style agents from both the…
Benchmarks are essential for quantitatively tracking progress in AI. As AI agents become increasingly capable, researchers and practitioners have introduced agentic benchmarks to evaluate agents on complex, real-world tasks. These…
With the development of foundation model (FM), agentic AI systems are getting more attention, yet their inherent issues like hallucination and poor reasoning, coupled with the frequent ad-hoc nature of system design, lead to unreliable and…
Large Language Model (LLM)-based coding agents show promise in automating software development tasks, yet they frequently fail in ways that are difficult for developers to understand and debug. While general-purpose LLMs like GPT can…
We introduce AuditBench, an alignment auditing benchmark. AuditBench consists of 56 language models with implanted hidden behaviors. Each model has one of 14 concerning behaviors--such as sycophantic deference, opposition to AI regulation,…
AI agents are increasingly used to solve complex, multi-step tasks, but existing multi-agent frameworks remain brittle as workflows grow in scale and depth. Small errors at intermediate stages can propagate through agent interactions, while…
Real-world requests to AI agents are fundamentally underspecified. Natural human communication relies on shared context and unstated constraints that speakers expect listeners to infer. Current agentic benchmarks test explicit…
As the focus in LLM-based coding shifts from static single-step code generation to multi-step agentic interaction with tools and environments, understanding which tasks will challenge agents and why becomes increasingly difficult. This is…
Over the last decade, explainable AI has primarily focused on interpreting individual model predictions, producing post-hoc explanations that relate inputs to outputs under a fixed decision structure. Recent advances in large language…
Testing conversational AI systems at scale across diverse domains necessitates realistic and diverse user interactions capturing a wide array of behavioral patterns. We present a novel multi-agent framework for realistic, explainable human…
Agent benchmarks have become the de facto measure of frontier AI competence, guiding model selection, investment, and deployment. However, reward hacking, where agents maximize a score without performing the intended task, emerges…
While individual components of agentic architectures have been studied in isolation, there remains limited empirical understanding of how different design dimensions interact within complex multi-agent systems. This study aims to address…
Agentic AI marks an important transition from single-step generative models to systems capable of reasoning, planning, acting, and adapting over long-lasting tasks. By integrating memory, tool use, and iterative decision cycles, these…
Fine-tuning large language models for code editing has typically relied on mining commits and pull requests. The working hypothesis has been that commit messages describe human intent in natural language, and patches to code describe the…
When an autonomous agent behaves undesirably, including failure to complete a task, it can be difficult to determine whether the behavior is due to a systemic agent error, such as flaws in the model or policy, or an environment error, where…
Agentic systems for business process automation often require compliance with policies governing conditional updates to the system state. Evaluation of policy adherence in LLM-based agentic workflows is typically performed by comparing the…
The rapid integration of Large Language Models (LLMs) into high-stakes domains necessitates reliable safety and compliance evaluation. However, existing static benchmarks are ill-equipped to address the dynamic nature of AI risks and…
The rise of agentic AI systems, where agents collaborate to perform diverse tasks, poses new challenges with observing, analyzing and optimizing their behavior. Traditional evaluation and benchmarking approaches struggle to handle the…
Despite rapid progress in building conversational AI agents, robustness is still largely untested. Small shifts in user behavior, such as being more impatient, incoherent, or skeptical, can cause sharp drops in agent performance, revealing…