Rectifying Shortcut Behaviors in Preference-based Reward Learning

• URL: http://arxiv.org/abs/2510.19050v1
• Date: Tue, 21 Oct 2025 20:08:32 GMT
• Title: Rectifying Shortcut Behaviors in Preference-based Reward Learning
• Authors: Wenqian Ye, Guangtao Zheng, Aidong Zhang,
• Abstract summary: In reinforcement learning, preference-based reward models play a central role in aligning large language models to human-aligned behavior.<n>Recent studies show that these models are prone to reward hacking and often fail to generalize well due to over-optimization.<n>We introduce a principled yet flexible approach to mitigate shortcut behaviors in preference-based reward learning.
• Score: 46.09046818725698
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: In reinforcement learning from human feedback, preference-based reward models play a central role in aligning large language models to human-aligned behavior. However, recent studies show that these models are prone to reward hacking and often fail to generalize well due to over-optimization. They achieve high reward scores by exploiting shortcuts, that is, exploiting spurious features (e.g., response verbosity, agreeable tone, or sycophancy) that correlate with human preference labels in the training data rather than genuinely reflecting the intended objectives. In this paper, instead of probing these issues one at a time, we take a broader view of the reward hacking problem as shortcut behaviors and introduce a principled yet flexible approach to mitigate shortcut behaviors in preference-based reward learning. Inspired by the invariant theory in the kernel perspective, we propose Preference-based Reward Invariance for Shortcut Mitigation (PRISM), which learns group-invariant kernels with feature maps in a closed-form learning objective. Experimental results in several benchmarks show that our method consistently improves the accuracy of the reward model on diverse out-of-distribution tasks and reduces the dependency on shortcuts in downstream policy models, establishing a robust framework for preference-based alignment.

Related papers

• Utility-inspired Reward Transformations Improve Reinforcement Learning Training of Language Models [6.472081755630166]
  We show how linear aggregation of rewards exhibits some vulnerabilities.<n>We propose a transformation of reward functions inspired by economic theory of utility functions.<n>We show that models trained with Inada-transformations score as more helpful while being less harmful.
  arXiv  Detail & Related papers  (2025-01-08T19:03:17Z)
• Self-supervised Preference Optimization: Enhance Your Language Model with Preference Degree Awareness [27.43137305486112]
  We propose a novel Self-supervised Preference Optimization (SPO) framework, which constructs a self-supervised preference degree loss combined with the alignment loss. The results demonstrate that SPO can be seamlessly integrated with existing preference optimization methods to achieve state-of-the-art performance.
  arXiv  Detail & Related papers  (2024-09-26T12:37:26Z)
• Personalizing Reinforcement Learning from Human Feedback with Variational Preference Learning [12.742158403867002]
  Reinforcement Learning from Human Feedback is a powerful paradigm for aligning foundation models to human values and preferences. Current RLHF techniques cannot account for the naturally occurring differences in individual human preferences across a diverse population. We develop a class of multimodal RLHF methods to address the need for pluralistic alignment.
  arXiv  Detail & Related papers  (2024-08-19T15:18:30Z)
• Regularizing Hidden States Enables Learning Generalizable Reward Model for LLMs [25.011675414622392]
  This study introduces a novel approach to enhance the reward model's generalization ability against distribution shifts. We retain the base model's language model head and incorporate a suite of text-generation losses to preserve the hidden states' text-generation capabilities. Our experimental results demonstrate that the introduced regularization technique markedly improves the accuracy of learned reward models.
  arXiv  Detail & Related papers  (2024-06-14T17:49:59Z)
• Reinforcement Learning from Diverse Human Preferences [68.4294547285359]
  This paper develops a method for crowd-sourcing preference labels and learning from diverse human preferences. The proposed method is tested on a variety of tasks in DMcontrol and Meta-world. It has shown consistent and significant improvements over existing preference-based RL algorithms when learning from diverse feedback.
  arXiv  Detail & Related papers  (2023-01-27T15:18:54Z)
• Reward Uncertainty for Exploration in Preference-based Reinforcement Learning [88.34958680436552]
  We present an exploration method specifically for preference-based reinforcement learning algorithms. Our main idea is to design an intrinsic reward by measuring the novelty based on learned reward. Our experiments show that exploration bonus from uncertainty in learned reward improves both feedback- and sample-efficiency of preference-based RL algorithms.
  arXiv  Detail & Related papers  (2022-05-24T23:22:10Z)
• SURF: Semi-supervised Reward Learning with Data Augmentation for Feedback-efficient Preference-based Reinforcement Learning [168.89470249446023]
  We present SURF, a semi-supervised reward learning framework that utilizes a large amount of unlabeled samples with data augmentation. In order to leverage unlabeled samples for reward learning, we infer pseudo-labels of the unlabeled samples based on the confidence of the preference predictor. Our experiments demonstrate that our approach significantly improves the feedback-efficiency of the preference-based method on a variety of locomotion and robotic manipulation tasks.
  arXiv  Detail & Related papers  (2022-03-18T16:50:38Z)
• Generative Adversarial Reward Learning for Generalized Behavior Tendency Inference [71.11416263370823]
  We propose a generative inverse reinforcement learning for user behavioral preference modelling. Our model can automatically learn the rewards from user's actions based on discriminative actor-critic network and Wasserstein GAN.
  arXiv  Detail & Related papers  (2021-05-03T13:14:25Z)
• Learning Diverse Representations for Fast Adaptation to Distribution Shift [78.83747601814669]
  We present a method for learning multiple models, incorporating an objective that pressures each to learn a distinct way to solve the task. We demonstrate our framework's ability to facilitate rapid adaptation to distribution shift.
  arXiv  Detail & Related papers  (2020-06-12T12:23:50Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.