Related papers: SpecBench: Measuring Reward Hacking in Long-Horizo…
Agent systems often decompose a task across multiple roles, but these roles are typically specified by prompts rather than enforced by access controls. Without enforcement, a team pass rate can mask whether agents actually coordinated or…
Repository-level code agents have shown strong promise in real-world feature addition tasks, making reliable evaluation of their capabilities increasingly important. However, existing benchmarks primarily evaluate these agents as black…
Beyond scratch coding, exploiting large-scale code repositories (e.g., GitHub) for practical tasks is vital in real-world software development, yet current benchmarks rarely evaluate code agents in such authentic, workflow-driven scenarios.…
The rapid advancement of multimodal large language models has enabled agents to operate mobile devices by directly interacting with graphical user interfaces, opening new possibilities for mobile automation. However, real-world mobile tasks…
Large language models (LLMs) now support automated software security tasks, including vulnerability discovery and proof-of-concept (PoC) generation. Existing benchmarks do not faithfully evaluate LLMs in real-world bug hunting scenarios…
Generative AI agents, software systems powered by Large Language Models (LLMs), are emerging as a promising approach to automate cybersecurity tasks. Among the others, penetration testing is a challenging field due to the task complexity…
Reward hacking--where agents exploit flaws in imperfect reward functions rather than performing tasks as intended--poses risks for AI alignment. Reward hacking has been observed in real training runs, with coding agents learning to…
Reward hacking in code generation, where models exploit evaluation loopholes to obtain full reward without correctly solving the tasks, poses a critical challenge for Reinforcement Learning (RL) and the deployment of reasoning models.…
Robotic manipulation policies often degrade over extended horizons, yet existing benchmarks provide limited insight into why such failures occur. Most prior benchmarks are either simulation-based or report aggregate success, making it…
Benchmarks for large language models (LLMs) have predominantly assessed short-horizon, localized reasoning. Existing long-horizon suites (e.g. SWE-bench) rely on manually curated issues, so expanding or tuning difficulty demands expensive…
Modern coding scaffolds turn LLMs into capable software agents, but their ability to follow scaffold-specified instructions remains under-examined, especially when constraints are heterogeneous and persist across interactions. To fill this…
Reward hacking, where a reasoning model exploits loopholes in a reward function to achieve high rewards without solving the intended task, poses a significant threat. This behavior may be explicit, i.e. verbalized in the model's…
Turning ideas into full software projects from scratch has become a popular use case for language models. Agents are being deployed to seed, maintain, and grow codebases over extended periods with minimal human oversight. Such settings…
Coding agents are increasingly used as iterative development partners, but most benchmarks still evaluate one specification followed by one final assessment. This leaves out a basic question: can an agent keep its own codebase working as…
Since autonomous coding agents generate complex behaviors at high-volume, we may want to use other LLMs to monitor actions to reduce the risk from dangerous misaligned behavior. To better understand the limitations of such monitors against…
Benchmark hacking refers to tuning a machine learning model to score highly on certain evaluation criteria without improving true generalization or faithfully solving the intended problem. We study this phenomenon in a generic machine…
As language agents increasingly automate critical tasks, their ability to follow domain-specific standard operating procedures (SOPs), policies, and constraints when taking actions and making tool calls becomes essential yet remains…
As Multimodal Large Language Models (MLLMs) advance, multimodal agents show promise in real-world tasks like web navigation and embodied intelligence. However, due to limitations in a lack of external feedback, these agents struggle with…
We introduce SWE-Bench Pro, a substantially more challenging benchmark that builds upon the best practices of SWE-BENCH [25], but is explicitly designed to capture realistic, complex, enterprise-level problems beyond the scope of SWE-BENCH.…
Frontier coding agents solve complex tasks when given complete context but collapse when specifications are incomplete or ambiguous. The bottleneck is not raw capability, but judgment: knowing when to act autonomously and when to ask for…