相关论文: Retrying vs Resampling in AI Control
Malware detection increasingly relies on AI systems that integrate signature-based detection with machine learning. However, these components are typically developed and combined in isolation, missing opportunities to reduce data complexity…
Privacy-preserving inference in edge computing paradigms encourages the users of machine-learning services to locally run a model on their private input and only share the models outputs for a target task with the server. We study how a…
Large Language Models (LLMs) perform impressively well in various applications. However, the potential for misuse of these models in activities such as plagiarism, generating fake news, and spamming has raised concern about their…
Recent reasoning-focused language models such as DeepSeek R1 and OpenAI o1 have demonstrated strong performance on structured reasoning benchmarks including GSM8K, MATH, and multi-hop question answering tasks. However, their performance…
Powered by the increasing predictive capabilities of machine learning algorithms, artificial intelligence (AI) systems have the potential to overrule human mistakes in many settings. We provide the first field evidence that the use of AI…
Information-flow control mechanisms are difficult both to design and to prove correct. To reduce the time wasted on doomed proof attempts due to broken definitions, we advocate modern random testing techniques for finding counterexamples…
Powerful new frontier AI technologies are bringing many benefits to society but at the same time bring new risks. AI developers and regulators are therefore seeking ways to assure the safety of such systems, and one promising method under…
Importance sampling (IS) is a popular technique in off-policy evaluation, which re-weights the return of trajectories in the replay buffer to boost sample efficiency. However, training with IS can be unstable and previous attempts to…
In learning problems, the noise inherent to the task at hand hinders the possibility to infer without a certain degree of uncertainty. Quantifying this uncertainty, regardless of its wide use, assumes high relevance for security-sensitive…
For many compiled languages, source-level types are erased very early in the compilation process. As a result, further compiler passes may convert type-safe source into type-unsafe machine code. Type-unsafe idioms in the original source and…
Large language models are increasingly solving tasks that are commonly believed to require human-level reasoning ability. However, these models still perform very poorly on benchmarks of general intelligence such as the Abstraction and…
Background: Screening trials require large sample sizes and long time-horizons to demonstrate mortality reductions. We recently proposed increasing statistical power by testing stored control-arm specimens, called the Intended Effect (IE)…
AI control protocols serve as a defense mechanism to stop untrusted LLM agents from causing harm in autonomous settings. Prior work treats this as a security problem, stress testing with exploits that use the deployment context to subtly…
Detecting semantic backdoors in classification models--where some classes can be activated by certain natural, but out-of-distribution inputs--is an important problem that has received relatively little attention. Semantic backdoors are…
Practitioners have reported a directional pattern in AI-assisted code generation: AI-generated code tends to fail quietly, preserving the appearance of functionality while degrading or concealing guarantees. This paper introduces the…
It is not an exaggeration to say that the recent progress in artificial intelligence technology depends on large-scale and high-quality data. Simultaneously, a prevalent issue exists everywhere: the budget for data labeling is constrained.…
Rerankers, typically cross-encoders, are computationally intensive but are frequently used because they are widely assumed to outperform cheaper initial IR systems. We challenge this assumption by measuring reranker performance for full…
Deep neural networks usually require large labeled datasets for training to achieve state-of-the-art performance in many tasks, such as image classification and natural language processing. Although a lot of data is created each day by…
Early identification of security issues in software development is vital to minimize their unanticipated impacts. Code review is a widely used manual analysis method that aims to uncover security issues along with other coding issues in…
AI is increasingly being used to assist fraud and cybercrime. However, it is unclear the extent to which current large language models can provide useful information for complex criminal activity. Working with law enforcement and policy…