English
Related papers

Related papers: Inference-Time Reward Hacking in Large Language Mo…

200 papers

Reward models learned from human preferences are central to aligning large language models (LLMs) via reinforcement learning from human feedback, yet they are often vulnerable to reward hacking due to noisy annotations and systematic biases…

Machine Learning · Computer Science 2026-02-12 Zhibin Duan , Guowei Rong , Zhuo Li , Bo Chen , Mingyuan Zhou , Dandan Guo

Direct alignment algorithms have proven an effective step for aligning language models to human-desired behaviors. Current variants of the Direct Preference Optimization objective have focused on a strict setting where all tokens are…

Computation and Language · Computer Science 2025-11-03 Fenia Christopoulou , Ronald Cardenas , Gerasimos Lampouras , Haitham Bou-Ammar , Jun Wang

The correct specification of reward models is a well-known challenge in reinforcement learning. Hand-crafted reward functions often lead to inefficient or suboptimal policies and may not be aligned with user values. Reinforcement learning…

Artificial Intelligence · Computer Science 2024-10-24 Muhan Lin , Shuyang Shi , Yue Guo , Behdad Chalaki , Vaishnav Tadiparthi , Ehsan Moradi Pari , Simon Stepputtis , Joseph Campbell , Katia Sycara

Specifying reward functions for complex tasks like object manipulation or driving is challenging to do by hand. Reward learning seeks to address this by learning a reward model using human feedback on selected query policies. This shifts…

Machine Learning · Computer Science 2023-02-27 Kush Bhatia , Wenshuo Guo , Jacob Steinhardt

Designing reward functions for efficiently guiding reinforcement learning (RL) agents toward specific behaviors is a complex task. This is challenging since it requires the identification of reward structures that are not sparse and that…

Machine Learning · Computer Science 2023-11-01 Dhawal Gupta , Yash Chandak , Scott M. Jordan , Philip S. Thomas , Bruno Castro da Silva

A safe and trustworthy use of Large Language Models (LLMs) requires an accurate expression of confidence in their answers. We propose a novel Reinforcement Learning approach that allows to directly fine-tune LLMs to express calibrated…

Computation and Language · Computer Science 2026-03-03 David Bani-Harouni , Chantal Pellegrini , Paul Stangel , Ege Özsoy , Kamilia Zaripova , Nassir Navab , Matthias Keicher

Language model calibration refers to the alignment between the confidence of the model and the actual performance of its responses. While previous studies point out the overconfidence phenomenon in Large Language Models (LLMs) and show that…

Computation and Language · Computer Science 2025-03-04 Jixuan Leng , Chengsong Huang , Banghua Zhu , Jiaxin Huang

The alignment of large language models (LLMs) with human preferences is commonly achieved through Reinforcement Learning from Human Feedback (RLHF). Direct Preference Optimization (DPO) simplified this paradigm by establishing a direct…

Machine Learning · Computer Science 2025-09-26 Yuandong Tan

This work tackles the problem of overoptimization in reinforcement learning from human feedback (RLHF), a prevalent technique for aligning models with human preferences. RLHF relies on reward or preference models trained on \emph{fixed…

Machine Learning · Computer Science 2025-03-11 Dhawal Gupta , Adam Fisch , Christoph Dann , Alekh Agarwal

This paper concerns the problem of aligning samples from large language models to human preferences using best-of-$n$ sampling, where we draw $n$ samples, rank them, and return the best one. We consider two fundamental problems. First: what…

Computation and Language · Computer Science 2024-11-05 Lin Gui , Cristina Gârbacea , Victor Veitch

While large language models demonstrate remarkable capabilities, they often present challenges in terms of safety, alignment with human values, and stability during training. Here, we focus on two prevalent methods used to align these…

Computation and Language · Computer Science 2023-10-26 Gabriel Mukobi , Peter Chatain , Su Fong , Robert Windesheim , Gitta Kutyniok , Kush Bhatia , Silas Alberti

Reinforcement Learning from Human Feedback aligns the outputs of Large Language Models with human values and preferences. Central to this process is the reward model (RM), which translates human feedback into training signals for optimising…

Artificial Intelligence · Computer Science 2025-09-23 Zeyu Huang , Zihan Qiu , Zili Wang , Edoardo M. Ponti , Ivan Titov

We show that learning algorithms satisfying a $\textit{low approximate regret}$ property experience fast convergence to approximate optimality in a large class of repeated games. Our property, which simply requires that each learner has…

Computer Science and Game Theory · Computer Science 2016-12-19 Dylan J. Foster , Zhiyuan Li , Thodoris Lykouris , Karthik Sridharan , Eva Tardos

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

Group Relative Policy Optimization (GRPO) has been shown to be an effective algorithm when an accurate reward model is available. However, such a highly reliable reward model is not available in many real-world tasks. In this paper, we…

Machine Learning · Computer Science 2026-01-12 Yuki Ichihara , Yuu Jinnai , Tetsuro Morimura , Mitsuki Sakamoto , Ryota Mitsuhashi , Eiji Uchibe

The safe and effective deployment of Large Language Models (LLMs) involves a critical step called alignment, which ensures that the model's responses are in accordance with human preferences. Prevalent alignment techniques, such as DPO, PPO…

Computation and Language · Computer Science 2024-11-04 Hanshi Sun , Momin Haider , Ruiqi Zhang , Huitao Yang , Jiahao Qiu , Ming Yin , Mengdi Wang , Peter Bartlett , Andrea Zanette

Reinforcement learning (RL) trained language model agents with tool access are increasingly deployed in coding assistants, research tools, and autonomous systems. We introduce the Reward Hacking Benchmark (RHB), a suite of multi-step tasks…

Machine Learning · Computer Science 2026-05-06 Kunvar Thaman

Games are challenging for Reinforcement Learning~(RL) agents due to their reward-sparsity, as rewards are only obtainable after long sequences of deliberate actions. Intrinsic Motivation~(IM) methods -- which introduce exploration rewards…

Artificial Intelligence · Computer Science 2025-07-29 Leonardo Villalobos-Arias , Grant Forbes , Jianxun Wang , David L Roberts , Arnav Jhala

Reward models (RMs) used in reinforcement learning from human feedback (RLHF) are vulnerable to reward hacking: as the policy maximizes a learned proxy reward, true quality plateaus or degrades. We make the assumption that reward hacking is…

Machine Learning · Computer Science 2026-04-06 Shinnosuke Ono , Johannes Ackermann , Soichiro Nishimori , Takashi Ishida , Masashi Sugiyama

Allocating more computation during inference time (test-time scaling) improves language model performance, especially for reasoning tasks. However, popular methods like Best-of-$N$ sampling often show diminishing returns as $N$ increases.…

Machine Learning · Computer Science 2025-10-20 Yung-Chen Tang , Pin-Yu Chen , Andrea Cavallaro
‹ Prev 1 3 4 5 6 7 10 Next ›