English
Related papers

Related papers: Detecting Proxy Gaming in RL and LLM Alignment via…

200 papers

During initial iterations of training in most Reinforcement Learning (RL) algorithms, agents perform a significant number of random exploratory steps. In the real world, this can limit the practicality of these algorithms as it can lead to…

Machine Learning · Computer Science 2022-10-17 Ashish Kumar Jayant , Shalabh Bhatnagar

We consider simple stochastic games with terminal-node rewards and multiple players, who have differing perceptions of risk. Specifically, we study risk-sensitive equilibria (RSEs), where no player can improve their perceived reward --…

Computer Science and Game Theory · Computer Science 2025-02-11 Léonard Brice , Thomas Henzinger , K. S. Thejaswini

Mitigating reward hacking--where AI systems misbehave due to flaws or misspecifications in their learning objectives--remains a key challenge in constructing capable and aligned models. We show that we can monitor a frontier reasoning…

Artificial Intelligence · Computer Science 2025-03-18 Bowen Baker , Joost Huizinga , Leo Gao , Zehao Dou , Melody Y. Guan , Aleksander Madry , Wojciech Zaremba , Jakub Pachocki , David Farhi

In cybersecurity, security analysts constantly face the challenge of mitigating newly discovered vulnerabilities in real-time, with over 300,000 vulnerabilities identified since 1999. The sheer volume of known vulnerabilities complicates…

Cryptography and Security · Computer Science 2026-01-26 Reza Fayyazi , Stella Hoyos Trueba , Michael Zuzak , Shanchieh Jay Yang

Solving complex mathematical problems via system-2 reasoning is a natural human skill, yet it remains a significant challenge for current large language models (LLMs). We identify the scarcity of deliberate multi-step reasoning data as a…

Artificial Intelligence · Computer Science 2024-12-25 Huanqia Cai , Yijun Yang , Zhifeng Li

As jailbreaks, adversarially crafted inputs that bypass safety constraints, continue to be discovered in Large Language Models, practitioners increasingly rely on fine-tuning as a defensive strategy. Yet the theoretical foundations…

Machine Learning · Computer Science 2026-05-20 Paul Wang , Jade Garcia-Bourrée , Anne-Marie Kermarrec , Vincent Corruble

Structural equation models (SEMs) are widely used in sciences, ranging from economics to psychology, to uncover causal relationships underlying a complex system under consideration and estimate structural parameters of interest. We study…

Machine Learning · Statistics 2020-10-21 Luofeng Liao , You-Lin Chen , Zhuoran Yang , Bo Dai , Zhaoran Wang , Mladen Kolar

AI agents are entering high-risk production settings, where they use tools, retain context, follow policies, handle private data, and interact with users over multiple turns. Yet many evaluation methods still judge isolated outputs or…

Multiagent Systems · Computer Science 2026-05-26 Fouad Bousetouane

Large language models are increasingly deployed as specialized agents that plan, call tools, and take actions over extended horizons. Yet many existing evaluations assume a "clean interface" where dynamics are specified and stable, tools…

Computation and Language · Computer Science 2026-02-04 Pouya Pezeshkpour , Estevam Hruschka

We propose using reinforcement learning to address the challenges of discovering microarchitectural vulnerabilities, such as Spectre and Meltdown, which exploit subtle interactions in modern processors. Traditional methods like random…

Cryptography and Security · Computer Science 2025-02-21 M. Caner Tol , Kemal Derya , Berk Sunar

Large Language Models (LLMs) are increasingly attracting attention in various applications. Nonetheless, there is a growing concern as some users attempt to exploit these models for malicious purposes, including the synthesis of controlled…

Artificial Intelligence · Computer Science 2025-08-26 Chongwen Zhao , Zhihao Dou , Kaizhu Huang

Reinforcement learning (RL) has enabled complex reasoning abilities in large language models (LLMs). However, most RL algorithms suffer from performance saturation, preventing continued gains as RL training scales. This problem can be…

Machine Learning · Computer Science 2026-05-12 Bolian Li , Yifan Wang , Yi Ding , Anamika Lochab , Ananth Grama , Ruqi Zhang

We introduce ProofGrid, a benchmark suite for evaluating LLM reasoning through machine-checkable proofs rather than final answers alone. ProofGrid contains 15 tasks spanning proof writing, proof checking, proof masking, and proof…

Logic in Computer Science · Computer Science 2026-05-14 Konstantine Arkoudas , Serafim Batzoglou

Machine unlearning aims to remove specific data influences from trained models, a capability essential for adhering to copyright laws and ensuring AI safety. Current unlearning metrics typically measure success by monitoring the model's…

Cryptography and Security · Computer Science 2025-12-24 Hengrui Jia , Taoran Li , Jonas Guan , Varun Chandrasekaran

Large language model pretraining is compute-intensive, yet many tokens contribute marginally to learning, resulting in inefficiency. We introduce Efficient Selective Language Modeling (ESLM), a risk-aware algorithm that improves training…

Machine Learning · Computer Science 2025-05-27 Melis Ilayda Bal , Volkan Cevher , Michael Muehlebach

Game theory provides the gold standard for analyzing adversarial engagements, offering strong optimality guarantees. However, these guarantees often become brittle when assumptions such as perfect information are violated. Reinforcement…

Machine Learning · Computer Science 2026-03-18 Goutam Das , Michael Dorothy , Kyle Volle , Daigo Shishika

The Extreme Learning Machine (ELM) is a growing statistical technique widely applied to regression problems. In essence, ELMs are single-layer neural networks where the hidden layer weights are randomly sampled from a specific distribution,…

Machine Learning · Statistics 2025-07-31 Daniela De Canditiis , Fabiano Veglianti

The safety and alignment of Large Language Models (LLMs) are critical for their responsible deployment. Current evaluation methods predominantly focus on identifying and preventing overtly harmful outputs. However, they often fail to…

Fine-tuning lets practitioners repurpose aligned large language models (LLMs) for new domains, yet recent work reveals emergent misalignment (EMA): Even a small, domain-specific fine-tune can induce harmful behaviors far outside the target…

Machine Learning · Computer Science 2026-03-06 David Kaczér , Magnus Jørgenvåg , Clemens Vetter , Esha Afzal , Robin Haselhorst , Lucie Flek , Florian Mai

While video large language models (Video-LLMs) excel in understanding slow-paced, real-world egocentric videos, their capabilities in high-velocity, information-dense virtual environments remain under-explored. Existing benchmarks focus on…

Computer Vision and Pattern Recognition · Computer Science 2026-04-21 Jianzhe Ma , Zhonghao Cao , Shangkui Chen , Yichen Xu , Wenxuan Wang , Qin Jin