PREFINE: Preference-Based Implicit Reward and Cost Fine-Tuning for Safety Alignment

• URL: http://arxiv.org/abs/2605.21225v1
• Date: Wed, 20 May 2026 14:19:45 GMT
• Title: PREFINE: Preference-Based Implicit Reward and Cost Fine-Tuning for Safety Alignment
• Abstract summary: We introduce PREFINE: Preference-based Implicit Reward and Cost Fine-Tuning for Safety Alignment.<n> PREFINE produces policies that achieve low-cost, high-reward performance with significantly improved data and computational efficiency.
• Score: 10.476492988819372
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: We address the problem of making a pre-trained reinforcement learning (RL) policy safety-aware by incorporating cost constraints without retraining it from scratch. While costs could be numerically encoded, we assume a more general setting is when costs are provided as preferences. Given a reward-optimized policy and a small dataset of preferred (low-cost) and dispreferred (high-cost) trajectories, our goal is to fine-tune the policy to generate low-cost behaviors while retaining high rewards. Unlike standard RLHF in language models, where preferences are defined over responses to the same prompt, our setting involves trajectory-level preferences in continuous control environments. We introduce PREFINE: Preference-based Implicit Reward and Cost Fine-Tuning for Safety Alignment which is a preference-based fine-tuning method that adapts Direct Preference Optimization (DPO), which is now widely used for LLM fine-tuning, to the sequential decision making setting. PREFINE constructs policy-sampled counterfactual trajectories to establish meaningful preference contrasts and jointly optimizes for reward retention and safety alignment. Empirically, PREFINE reduces constraint violations and catastrophic failures by over 60% while maintaining original reward behavior. PREFINE produces policies that achieve low-cost, high-reward performance with significantly improved data and computational efficiency compared to full offline RL or imitation learning, bridging preference alignment and safe policy adaptation in continuous domains.

Related papers

• Ratio-Variance Regularized Policy Optimization [64.95520246570446]
  We show that explicitly constraining the policy ratio variance provides a principled local approximation to trust-region constraints.<n>We introduce $bf R2bf VPO$ (Ratio-Variance Regularized Policy Optimization), which implements this constraint via a primal-dual optimization framework.
  arXiv  Detail & Related papers  (2026-05-26T09:53:42Z)
• Robust Policy Optimization to Prevent Catastrophic Forgetting [29.514746370429965]
  Large language models are commonly trained through multi-stage post-training.<n>Small downstream updates can compromise earlier learned behaviors.<n>This suggests standard RLHF objectives do not guarantee robustness to future adaptation.
  arXiv  Detail & Related papers  (2026-02-09T15:50:05Z)
• Conditional Sequence Modeling for Safe Reinforcement Learning [8.858563919623082]
  offline safe reinforcement learning aims to learn policies from a fixed dataset while maximizing performance under cumulative cost constraints.<n>Most existing offline safe RL methods are trained under a pre-specified threshold.<n>We propose RCDT, a CSM-based method that supports zero-shot deployment across multiple cost thresholds within a single trained policy.
  arXiv  Detail & Related papers  (2026-02-09T12:22:57Z)
• Primal-Dual Direct Preference Optimization for Constrained LLM Alignment [16.080857375857697]
  We study the problem of constrained alignment in Large Language Models (LLMs)<n>We propose a novel primal-dual DPO approach, which first trains a model using standard DPO on reward preference data.<n>We establish rigorous theoretical guarantees on the suboptimality and constraint violation of the output policy.
  arXiv  Detail & Related papers  (2025-10-07T09:10:35Z)
• Stabilizing Policy Gradients for Sample-Efficient Reinforcement Learning in LLM Reasoning [77.92320830700797]
  Reinforcement Learning has played a central role in enabling reasoning capabilities of Large Language Models.<n>We propose a tractable computational framework that tracks and leverages curvature information during policy updates.<n>The algorithm, Curvature-Aware Policy Optimization (CAPO), identifies samples that contribute to unstable updates and masks them out.
  arXiv  Detail & Related papers  (2025-10-01T12:29:32Z)
• On Symmetric Losses for Robust Policy Optimization with Noisy Preferences [55.8615920580824]
  This work focuses on reward modeling, a core component in reinforcement learning from human feedback.<n>We propose a principled framework for robust policy optimization under noisy preferences.<n>We prove that symmetric losses enable successful policy optimization even under noisy labels.
  arXiv  Detail & Related papers  (2025-05-30T15:30:43Z)
• AlphaDPO: Adaptive Reward Margin for Direct Preference Optimization [45.46582930202524]
  $alpha$-DPO is an adaptive preference optimization algorithm for large language models.<n>It balances the policy model and the reference model to achieve personalized reward margins.<n>It consistently outperforms DPO and SimPO across various model settings.
  arXiv  Detail & Related papers  (2024-10-14T04:29:57Z)
• Value-Incentivized Preference Optimization: A Unified Approach to Online and Offline RLHF [80.32171988565999]
  We introduce a unified approach to online and offline RLHF -- value-incentivized preference optimization (VPO)<n>VPO regularizes the maximum-likelihood estimate of the reward function with the corresponding value function.<n>Experiments on text summarization and dialog verify the practicality and effectiveness of VPO.
  arXiv  Detail & Related papers  (2024-05-29T17:51:42Z)
• Provably Mitigating Overoptimization in RLHF: Your SFT Loss is Implicitly an Adversarial Regularizer [52.09480867526656]
  We identify the source of misalignment as a form of distributional shift and uncertainty in learning human preferences.<n>To mitigate overoptimization, we first propose a theoretical algorithm that chooses the best policy for an adversarially chosen reward model.<n>Using the equivalence between reward models and the corresponding optimal policy, the algorithm features a simple objective that combines a preference optimization loss and a supervised learning loss.
  arXiv  Detail & Related papers  (2024-05-26T05:38:50Z)
• COptiDICE: Offline Constrained Reinforcement Learning via Stationary Distribution Correction Estimation [73.17078343706909]
  offline constrained reinforcement learning (RL) problem, in which the agent aims to compute a policy that maximizes expected return while satisfying given cost constraints, learning only from a pre-collected dataset. We present an offline constrained RL algorithm that optimize the policy in the space of the stationary distribution. Our algorithm, COptiDICE, directly estimates the stationary distribution corrections of the optimal policy with respect to returns, while constraining the cost upper bound, with the goal of yielding a cost-conservative policy for actual constraint satisfaction.
  arXiv  Detail & Related papers  (2022-04-19T15:55:47Z)
• First Order Constrained Optimization in Policy Space [19.00289722198614]
  We propose a novel approach called First Order Constrained Optimization in Policy Space (FOCOPS) FOCOPS maximizes an agent's overall reward while ensuring the agent satisfies a set of cost constraints. We provide empirical evidence that our simple approach achieves better performance on a set of constrained robotics locomotive tasks.
  arXiv  Detail & Related papers  (2020-02-16T05:07:17Z)

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.